JPH1185802A - Computer-readable recording medium recording full-text search data and character string collation device - Google Patents

Abstract

(57) [Summary] [Problem] In a case where a special character is inserted into a character string to be searched to separate meanings in an information search processing device,
For example, if it is inserted at a ratio of one character to three characters of the input character string, the number of appearances of the character is enormous, and only the chain memory of the character is abnormally enormous, which overwhelms the chain memory. The purpose is to solve. SOLUTION: When a special character appears in a character string, character string conversion means 301, 304 converts the special character into a character which is not a search target by an adjacent character and outputs it to a two-character chain detector 302, 305. Two-character chain detecting means 30
2 extracts a two-character chain from the character string obtained by converting the special character string into a character that is not a search target, counts the number of appearances for each chain, and stores it in the two-character chain memory 3030. The comparator checks the presence or absence of a character chain using a two-character chain memory for the search target character string obtained by converting the special character according to the adjacent character.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used for a full-text search for searching for a character string described in a document in the field of information search processing, and inputs all the character strings contained in a registered document. The present invention relates to a computer-readable recording medium that records full-text search data for collating a character string, and a character string collating apparatus.

[0002]

2. Description of the Related Art FIG. 36 shows the configuration of a conventional character string collating apparatus, and FIG. 37 shows a conventional character collating method and the format of full-text search data. In FIG. 37 (a), 801
Is a character string "iroa-hani" input at the time of registration, 802 is a two-character chain "iro" to be registered first,
The two-letter chain "ro a" next to "2", and 804
The character chain “a wa” and 805 are the next two character chain “hani” after 804. Here, “a” means a special character inserted into a character string to separate meanings, for example, a space that frequently appears in Hangul.

In FIG. 37 (c), reference numeral 811 denotes a search character string "iro a ha ni" at the time of search, and 812 denotes a search character string which is searched first.
The character chain “Iro”, 813 is the next two-character sequence “roa” after 812, 814 is the next two-character sequence “8a” after 813,
815 is the two-letter chain “Hani” following 814.

In FIG. 37 (b), a two-character chain 802
The number of appearances n1 and n2 of “yes” and “ro”, the two-character chain 803 is the number of appearances n2 and n3 of “ro” and “a”, and the two-character chain 804 is the occurrence of “a” and “ha” Number of times n3, n4
, The two-character chain 805 is the number of appearances of “ha” and “ni”
Store n4 and n5.

FIG. 37 (d) shows a recording format in the case where a plurality of "iro" chains exist in the input character string.
That is, the sequence of "iro" is such that "iro" appeared at the n1st time and "ro" appeared at the n2th time, "iro" appeared at the nath time and "ro" appeared at the nbth time, ..., nxth "I" appeared in
It is composed of "ro" that appeared at the ny-th time.

At this time, in the conventional collation method, a two-character chain 802 corresponding to the “color” of the two-character chain 812 is detected.
A two-character chain 803 corresponding to “a” in the character chain 813
Is detected, and it is determined whether or not the number of appearances n2 of “ro” at this time matches.

If they match, then "a"
Is detected, and a two-character chain 804 corresponding to “a wa” of the two-character chain following the 813 is detected.
It is determined whether or not the number of appearances of the two matches.

If they match, then the "ha" detected at 804
Is detected, and a two-character chain 805 corresponding to “Hani” in the next two-character chain of 814 is detected.
It is determined whether or not the number of appearances of the two matches. If they match, it is determined that the character string 811 matches 801. As described above, the character strings are collated.

FIG. 36 shows a configuration of a conventional character string collation apparatus using the search data shown in FIG.

In FIG. 36, reference numeral 701 denotes a two-character chain 802, 803, 80 to be registered from a character string 801 to be registered.
2, a two-character chain detector 702 for detecting the two-character chains 802, 803, 804, 805 and the number of appearances of those characters; and 703, a two-character chain for searching from the character string 811 to be searched. Character chain 812,
A two-character chain detector that detects 813, 814, 815;
Reference numeral 704 denotes the two-character chain 812, 813, 814, and 815 detected by the two-character chain detector 703 in the two-character chain memory 702, and the number of appearances of the character preceding the detected two-character chain is detected by the last two characters. A comparator 705 determines whether or not the number of occurrences of the character after the character chain matches. The comparator 704 determines all character strings detected by the two-character chain detector 703 and matches the character strings. Is a control unit that determines

[0011]

However, in the above-described configuration, when a special character (such as a space in Hangul) is inserted into a character string input at the time of registration to separate meanings, for example, the character string may not be inserted. If one character is inserted into three characters of the input character string at a rate of one character, the number of appearances of the character becomes enormous, and only the chain memory of the character becomes abnormally enormous. Was. Further, for the same character chain, the process of extracting the chain based on the coincidence of the number of appearances is repeated many times, and there is a problem that it takes time.

The present invention solves the above-mentioned problems of the prior art, and forms a chain as a three-character chain sandwiching a specific special character used as a delimiter or the like, or forms the special character. Is changed to a character uniquely determined by the character adjacent to the character, or the character before the special character is replaced by the two characters of the character before and the character uniquely determined by the preceding character, and the special character The character after is changed to two characters, a character uniquely determined by the following character and the subsequent character, thereby reducing the number of special characters and avoiding an increase in the special character chain memory. It is an object of the present invention to efficiently perform a process of extracting a chain based on the matching of.

[0013]

According to the present invention, first, full-text search data is detected by detecting all two-character sequences consisting of characters other than special characters specified in advance for a character string to be searched. For each character chain, first data that records the number of appearances of the first character and the second character forming the two-character chain in the search target character string as a set, and special characters other than the special character in which a special character specified in advance is inserted All character chains consisting of two characters are detected, and a first character chain forming a character chain is detected for each of the character chains.
The second data, in which the number of appearances of the character and the second character in the search target character string are recorded as a set, is recorded separately from the first data and the second data. Are detected, and all character chains consisting of two characters other than the special character in which a special character specified in advance is inserted are detected, and the respective character chains are converted into the first data and the second data. The above problem is solved by searching data and comparing the number of appearances corresponding to the detected character chain to determine the presence or absence of a character chain as a search character string.

Second, the full-text search data is converted into characters that are not to be searched according to adjacent special characters of a special character specified in the character string to be searched. All 2 including non-target characters
A character chain is detected, and for each two-character chain, the number of appearances of the first character and the second character constituting the two-character chain in the search target character string is recorded as a set. According to the same rules applied to the data recorded on the recording medium, convert to characters that are not to be searched based on adjacent characters, and include characters that are not to be searched for the converted character string. The above problem is solved by detecting all the two-character chains, detecting the detected two-character chains from the storage medium, and comparing the corresponding appearance counts to determine the presence or absence of a character chain as a search character string. doing.

Third, the full-text search data is converted into two characters that are not to be searched according to adjacent special characters of a pre-specified special character of the character string to be searched. , And all the two-character chains including the characters that are not the target of the search are detected, and for each two-character chain, the number of appearances of the first character and the second character constituting the two-character chain in the search target character string is recorded as a set. In accordance with the same rule applied to the data recorded on the recording medium, the special character specified in advance in the character string is converted into two characters that are not searched based on adjacent characters, and the converted character string is obtained. On the other hand, all the two-character sequences including the two characters that are not searched are detected, the detected two-character sequences are detected from the storage medium, and the corresponding occurrence counts are compared to find a character sequence as a search character string. It solves the problem by determining the presence or absence.

Fourth, the full-text search data is searched for a character string to be searched, and a two-character chain is detected for every character. For the first character and the second character constituting the character chain of, the number of appearances of characters other than the special character specified in advance, and a fixed numerical value in the case of the special character specified in advance, are recorded as a set. In the data 3 and the search target character string, all three-character chains consisting of three characters in which a special character designated in advance is inserted are detected, and a three-character chain is formed for each three-character chain. The above problem is solved by recording and accessing fourth data in which the number of appearances of one character and the third character is recorded as a set.

Fifth, in the full-text search data, a two-character chain consisting of characters other than a special character specified in advance is detected for a character string to be searched, and a two-character chain is formed for every two-character chain. Fifth data in which the number of appearances of the first character and the second character in the search target character string are recorded as a set, and all three characters including a special character specified in advance with respect to the search target character string And a sixth character string recorded as a set of the number of appearances of the first character and the value 0 and a set of the value 0 and the number of appearances of the third character for each three-character chain. By recording and accessing data,
The above problem has been solved. Sixth, full-text search data
For the character string to be searched, all two-character strings consisting of characters other than the special characters specified in advance are detected, and the first and second character strings forming the two-character string are formed for each two-character string. In the seventh data recorded as a set of the number of appearances in and the search target character string, all three-character chains consisting of three characters in which a special character specified in advance is inserted are detected. The first character and the second character are converted from the special character of the second character constituting the three-character chain to the same character as the third character, and the number of appearances of the second character is set to the same value as the number of occurrences of the third character. A first two-character chain that generates two character chains of a second character and a third character, and forms a two-character chain for each two-character chain
The above problem is solved by recording and accessing eighth data in which the number of appearances of a character and a second character in a search target character string is recorded as a set.

Seventh, the search data used in the full-text search is detected. The search data detects all two-character chains in the search target character string, forms a two-character chain for every two-character chain, and In the case of a configuration of a character chain other than the special character designated in advance, ninth data in which a set of the number of appearances of the first character and the second character other than the special character designated in advance is recorded, In the case where the two-character chain is a character chain including a special character specified in advance, the number of applications for the first character or the second character corresponding to the special character is equal to or less than the maximum value of the number of appearances specified in advance. If the remainder is 0 or the remainder is 0, the maximum value, or the maximum value and the remainder, or if the first appearance count is less than the maximum value, the value less than or equal to the second maximum value is the first time Values and order are unique The have, in the tenth data storing and number of occurrences of the character is not a special character as a set, and, if the first character is a special character, the 10th set of data of the second
The tenth data sorted for each character type;
The above problem is solved by storing and accessing the data and the tenth data separately.

Eighth, the search data used in the full-text search detects all two-character sequences in the character string to be searched, and outputs a document number for each two-character sequence, the number of appearances for each character type of the two-character sequence, or an arbitrary number. Construct character chain data consisting of value pairs,
If the character chain data does not include a special character specified in advance, the size of storing the number of appearances of the first character and the number of occurrences of the second character is equal. The character chain data configured so that the size for storing the corresponding number of appearances is larger than the size for storing an arbitrary value corresponding to a character that does not include a special character. When a special character string specified in advance is included, the second character stores the specified value, and the first character of the next continuous character chain data is the value specified by the second character of the previous character chain data. The character chain data configured to be equal to can be stored and accessed, thereby solving the above problem.

Ninth, the full-text search data is obtained by detecting all two-character chains that do not include special characters in the character string to be searched, and forming characters for each two-character chain except for the special characters specified in advance. Regarding the first character and the second character constituting the chain, the characters other than the special character designated in advance are stored in the eleventh data recorded as a set of the number of occurrences thereof, and the special character designated in advance for the search target character string. As a two-character chain that straddles characters, the number of appearances of the first character type of the two-character chain preceding the special character and the number of appearances of the first character type of the two-character chain following the special character are recorded as a set. Twelfth
The above problem can be solved by recording and accessing twelfth data, which is a combination of the number of occurrences of the first character type of the two-character chain preceding the special character or the special character and the number of occurrences of the character immediately after the special character. Has been resolved.

Tenth, in the full-text search data, a two-character chain consisting of characters other than a special character specified in advance is detected for a character string to be searched, and a two-character chain is formed for each two-character chain. The first three data in which the appearance position of the first character or the second character in the search target character string is recorded as the appearance position of a two-character chain, and all the character chains into which a special character specified in advance is inserted are detected. In the fourteenth aspect, for each of the character chains, the appearance position of the first character constituting the character chain in the search target character string is recorded as the appearance position of the character chain.
Is recorded separately from the thirteenth data and the fourteenth data, and from the search character string, all two-character chains consisting of characters other than the special character specified in advance and the special character specified in advance are inserted. All of the detected character chains are searched, and the respective character chains are searched from the thirteenth data and the fourteenth data. By comparing the appearance positions corresponding to the detected character chains, the presence or absence of the character chain as the search character string is determined. The above problem has been solved by determining.

Eleventh, the full-text search data is converted into a non-search target character according to adjacent characters of a pre-specified special character of the search target character string. Detects all two-character chains including non-target characters, and sets the appearance position of the first character or the second character constituting the two-character chain in the search target character string as the appearance position of the two-character chain for each two-character chain Record, according to the same rules applied to the data recorded on the recording medium, the pre-specified special character of the search character string is converted into a character that is not a search target based on adjacent characters,
In the converted character string, all the two-character sequences including the characters that are not to be searched are detected, the detected two-character sequences are detected from the storage medium, and the corresponding occurrence positions are compared to find the search character. The above problem is solved by determining the presence or absence of a character chain as a column.

Twelfth, the full-text search data is converted into two characters that are not to be searched according to adjacent special characters before and after a special character specified in the character string to be searched. , All the two-character sequences including the characters not to be searched are detected.
The appearance position of the first character or the second character constituting the character chain in the search target character string is recorded as the appearance position of the two-character chain, and the special character designated in advance in the search character string is recorded on the recording medium. Are not searched for based on adjacent characters according to the same rules applied to
The character string is converted to a character string, all two-character sequences including two characters that are not to be searched are detected in the converted character string, the detected two-character sequence is detected from the storage medium, and a corresponding appearance position is detected. The above problem is solved by determining the presence or absence of a character chain as a search character string by comparing.
Thirteenth, the full-text search data is
Detects all two-character sequences consisting of characters other than the special characters specified in advance, and sets the appearance position of the first character or the second character constituting the two-character sequence in the search target character string to two characters for each of the two-character sequences. With respect to the data recorded as the appearance position of the chain and the search target character string, all three-character chains consisting of three characters in which a special character specified in advance is inserted are detected. The special character of the second character constituting the chain is converted into the same character as the third character, and the appearance position of the second character is set to the same value as the appearance position of the third character. Fifteenth data consisting of data that generates two two-character chains of the third character and records the appearance position of the first character or the second character of the two-character chain consisting of the first character and the second character of the three-character chain And the second and third characters of the three-character chain The first 16 of the data recording the occurrence position of the first character or the second character Ranaru 2 character chain, recording, by accessing solves the above problems.

Fourteenth, a computer-readable storage medium storing search data used for full-text search, wherein the search data detects a two-character chain and a character position in a search target character string, A set of a document number, a two-character chain, and a character position composed of a search target character string is configured as two-character chain information for each time. The positions of the special characters are excluded and numbering is performed in ascending order. Further, the storage medium storing character chain information for each first character of the character chain. The seventeenth data recorded as a set of one character and the second character, the character position of the first character at the character position excluding the special character, and the document number, and the character immediately before the special character and the special character Combined A two-character chain combining the eighteenth data consisting of a character chain, an arbitrary fixed value defined by the character type of a special character, and a document number, a special character and the character immediately following the special character, a special character In the two-character chain combining the nineteenth data consisting of the character position of the second character at the character position excluding the character and the document number, and the characters immediately before and after the special character, the character position excluding the special character 20th data composed of the character position of the first character and the document number, and the first character of the two-character chain of the seventeenth data, the eighteenth data, the nineteenth data, and the twentieth data , And stored as character chain information. For the seventeenth data, the eighteenth data, and the nineteenth data, when the first character of the two character chains is the same and the second character is a special character, Seventeenth data The first 18 data stored after recording a full-text search data by accessing solves the above problems.

Fifteenth, a computer-readable storage medium storing search data used for full-text search, wherein the search data detects a two-character chain and a character position with respect to a character string to be searched; A storage medium that stores a set of a document number, a two-character chain, and a character position, each of which is a search target character string, as two-character chain information, and stores character chain information for each first character of the character chain; The character position of the two-character chain information is numbered in ascending or descending order excluding the position of the special character specified in advance with respect to the beginning of the search target character string. The 21st data recorded as a set of one character and the second character, the character position of the first character, and the document number, and the character chain information including the special character are: , Just before special characters A character chain consisting of a character and a character immediately following a special character, a set of a character position immediately before the special character and a document number. The first and second characters of the character chain of the character chain information are special characters. When the first character or the second character of the character chain of the character chain information in the case where the character chain information does not include the character string information, the first character or the second character is separately recorded before or after the character chain information not including the special character. Data and character chain information including special characters, for the character immediately after the special character, the character chain that combines the character immediately after the special character and the character that follows that character, the character position immediately after the special character And a document number, and when the first character of the character chain of the character chain information matches the first character of the two-character chain when no special character is included, the character chain information without the special character is included. After or before The 23rd data configured to be recorded in each character and the character chain information including the special character include a character chain combining the character two characters before the special character and the character immediately after the special character, Full-text search data characterized in that the twenty-fourth data composed of the character position and the document number two characters before and the twenty-first data, the twenty-second data, the twenty-third data, and the twenty-fourth data are stored separately. By recording and accessing
The above problem has been solved.

Further, the character string collating device of the present invention comprises:
In addition, a computer-readable recording medium storing search data used for full-text search by the first means for solving the above-mentioned problem, and all two characters consisting of characters other than special characters specified in advance from a search character string First character chain detecting means for detecting a chain, and second character chain detecting means for detecting, from the search character string, all character chains consisting of two characters other than the special character in which a predetermined special character is inserted. And searching the two-character chain detected by the first character chain detecting means from the first data recorded on the recording medium, detecting the character chain detected by the second character chain detecting means, A comparison unit that determines whether or not there is a continuation of the character chain as the search character string by comparing the number of appearances corresponding to the detected character chain.

Secondly, a computer-readable recording medium storing search data used for full-text search by the second means, and a special character designated in advance as a search character string for data recorded on the recording medium. According to the same rule applied, a character string conversion unit that converts adjacent characters into characters that are not to be searched, and a character string that is converted by the character string conversion unit, including characters that are not to be searched. A two-character chain detecting means for detecting all the two-character chain, and a two-character chain detected by the two-character chain detecting means are detected from the recording medium, and the corresponding numbers of appearances are compared. Comparing means for judging the presence or absence of continuation of the character chain.

Third, a computer-readable recording medium on which search data used for full-text search by the third means is recorded, and a special character designated as a search character string is applied to data recorded on the recording medium. Not subject to search based on adjacent characters according to the same rule 2
A character string conversion means for converting to a character, a two-character chain detection means for detecting all two-character chains including two characters which are not to be searched from the character string converted by the character string conversion means, 2 detected by character chain detection means
A character string collating device comprising: a comparing unit that detects a character chain from the recording medium and determines whether there is a continuation of the character chain as a search character string by comparing the number of corresponding appearances.

Fourth, a computer-readable recording medium storing search data used for full-text search by the fourth means, and all two-character sequences consisting of characters other than special characters specified in advance from a search character string A first character chain detecting means for detecting all character chains consisting of three characters into which a special character designated in advance is inserted, from a search character string; and The two-character chain detected by the character chain detecting means is searched from the third data recorded on the recording medium, and the character chain detected by the third character chain detecting means is retrieved from the third data recorded on the recording medium. And a comparing unit that determines whether there is a continuation of the character chain as the search character string by comparing the number of appearances corresponding to the detected character chain by searching from the data of No. 4.

Fifth, a computer-readable recording medium storing search data used for full-text search by the fifth means, and all two-character sequences consisting of characters other than special characters specified in advance from a search character string A first character chain detecting means for detecting all character chains consisting of three characters into which a predetermined special character has been inserted, from a search character string; and The two-character chain detected by the character chain detecting means is searched from the fifth data recorded on the recording medium, and the character chain detected by the third character chain detecting means is retrieved from the fifth data recorded on the recording medium. And a comparing means for judging the presence or absence of continuation of the character chain as the search character string by comparing the number of appearances corresponding to the detected character chain by searching from the data of No. 6.

Sixth, a computer-readable recording medium storing search data used for full-text search by the sixth means, and a two-character chain consisting of characters other than special characters specified in advance from a search character string A first character chain detecting unit for detecting all three-character chains consisting of three characters into which a predetermined special character has been inserted, from a search character string; The two-character chain detected by the character chain detecting means is searched from the seventh data recorded on the recording medium, and the special character detected by the fourth character chain detecting means is converted to generate a two-character string.
Comparing means for searching for a character chain from the eighth data recorded on the recording medium and comparing the number of appearances corresponding to the detected character chain to determine whether or not there is a continuation of the character chain as the search character string; Is provided.

Seventh, a computer-readable storage medium storing search data used for full-text search by the seventh means, and a fifth method for detecting all two-character sequences that do not include special characters from a search character string A character chain detecting means, a sixth character chain detecting means for detecting all character chains including special characters from the search character string, and a two-character chain detected by the first character chain detecting means. When the search string is composed of a sixth character chain, a comparison unit that determines the presence or absence of continuation of the character chain as a search character string by comparing the number of appearances corresponding to the detected character chain. In the case of a two-character chain searched by the detection means, the presence or absence of the continuation of the character chain as a search character string is determined by comparing the number of occurrences of the detected character chain and the number of times of occurrence of the special character. Size It has comparing means for a configuration with.

Eighthly, a fifth character chain detecting means for detecting, from the storage medium, search data used for full-text search by the eighth means, and all two-character strings that do not include special characters from a search character string. If the two-character chain does not include a special character, the appearance of the second character of the detected character chain data is compared with the character chain data corresponding to the continuous character chain detected by the fifth character chain detection unit. By comparing the number of times and the number of appearances of the first character of the character chain data of the character chain following the character chain, a comparison unit that determines whether or not the character chain as a search character string is continuous is provided. ing.

Ninth, a first method for detecting, from the storage medium and the search character string, all two-character sequences consisting of characters other than the special characters specified in advance, using search data used for full-text search by ninth means. And a two-character chain before and after the special character and a two-character chain before and after the special character. A second character chain detecting means for detecting a character chain formed by combining the first character with the first character, or a character formed by combining the character of the first character of the two-character chain preceding the special character and the character immediately after the special character A second character chain detecting means for detecting a chain, and two characters detected by the first character chain detecting means.
The character chain is searched from the eleventh data or the twelfth data recorded on the storage medium. If the character chain is searched from the eleventh data, the character chain detected by the seventh character chain detecting means is searched. When the search is performed from the second data, the character chain detected by the first character chain detection means is searched, and the number of appearances corresponding to the detected character chain is compared, so that the character chain can be used as a search character string. And a comparing means for judging the presence or absence of the continuation.

Tenth, a computer-readable recording medium on which search data used for full-text search by the tenth means is recorded, and all two-byte characters consisting of characters other than special characters specified in advance from a search character string. A first character chain detecting means for detecting a character chain, a second character chain detecting means for detecting, from the search character string, all character chains of three characters into which a predetermined special character is inserted, The two-character chain detected by the first character-chain detecting means is searched from the thirteenth data recorded on the recording medium, and the character chain detected by the eighth character-chain detecting means is recorded on the recording medium. And a comparing unit that determines whether there is a continuation of the character chain as the search character string by comparing the appearance positions corresponding to the detected character chains by searching from the fourteenth data. To have.

Eleventh, a computer-readable recording medium recording search data used for full-text search by the eleventh means, and a special character designated in advance as a search character string are stored in the data recorded in the recording medium. A character string conversion unit that converts adjacent characters into characters that are not to be searched according to the same rule that has been applied, and detects all two-character chains from the search character string, including characters that are not to be searched. A character chain detecting unit and a two-character chain detected by the two-character chain detecting unit are searched from the data recorded on the recording medium, and an appearance position corresponding to the detected character chain is compared to form a search character string. And a comparing means for determining whether or not there is a continuation of the character chain.

Twelfth, a computer-readable recording medium storing search data used for full-text search by the twelfth means, and a special character designated in advance as a search character string for data recorded on the recording medium. A character string conversion unit that converts adjacent characters into two characters that are not to be searched according to the same rule that has been applied, and all character strings, including two characters that are not to be searched, are converted from a search character string to a character string. Two-character chain detecting means for detecting a two-character chain;
The two-character chain detected by the character chain detecting means is searched from the data recorded on the recording medium, and the appearance position corresponding to the detected character chain is compared to determine whether or not the character chain as a search character string is continuous. And a comparing means for determining

Thirteenth, a computer-readable recording medium storing search data used for full-text search by the thirteenth means, and a two-character chain consisting of characters other than special characters specified in advance from a search character string And detects all three-character chains consisting of three characters with a special character inserted in advance, and for each three-character chain, the second special character constituting the three-character chain is the same as the third character A first character chain detecting means for converting the character sequence into a character and detecting a two-character chain composed of a first character and a second character; and a second character sequence detecting a two-character chain composed of a second character and a third character of the three-character sequence. The two character chains detected by the ten character chain detecting means and the first character chain detecting means are searched from the fifteenth data recorded on the recording medium, and are detected by the second character chain detecting means. Special characters The two-character chain generated by the replacement is searched from the sixteenth data recorded on the recording medium, and by comparing the appearance positions corresponding to the detected character chain, the presence or absence of the continuation of the character chain as the search character string is determined. And a comparing means for determining

Fourteenthly, a computer-readable recording medium storing search data used for full-text search by the fourteenth means, and all two-character sequences except special characters and characters before and after the special characters are detected from a search character string. And a two-character chain consisting of the character immediately before and after the special character, a two-character chain consisting of the character immediately before the special character and the special character, and a special character and special character from the search character string. A twelfth character chain detecting means for detecting a two-character chain consisting of a character immediately following a character, and seventeenth data corresponding to the two-character chain detected by the eleventh character chain detecting means and a twelfth character chain detection From the 20th data corresponding to the character chain detected by the means, or the 21st data and the 17th data.
Comparing means for judging the presence or absence of continuation as a search character string by comparing the character position difference between the two data with the document number, and including a special character because the second data immediately follows the twentieth data A comparison means for determining whether or not there is a continuation as a search character string is provided.

Fifteenth, a computer-readable recording medium storing search data used for full-text search by the fifteenth means, and all two-character sequences excluding characters before and after a special character are detected from a search character string. A thirteenth character chain detecting means for detecting, from the search character string, a character immediately before and after the special character as a character chain in the case of a search character string sandwiching a special character; The character is marked as the character immediately following the special character. If the beginning of the search string is a special character, the character immediately after the special character and the next character are detected as a character chain, and the first character of the character chain is If the special character appears after the third character from the beginning of the search string, the character two characters before the special character and the character immediately after the special character are detected as a character chain. Or The second character of the character chain is marked as the character immediately after the special character, and the character immediately after the special character and the next character are detected as a character chain, and the first character of the character chain is immediately after the special character. The fourteenth character chain detecting means for marking the character as
In the case of a two-character chain detected by the character chain detecting means, the character position corresponding to the detected character chain is compared with the document number to determine the presence or absence of the continuation of the character chain as the search character string. And the search string are the first
In the case of a two-character chain searched by the character chain detecting means of No. 4, a character string as a search character string is determined from the character chain and the document number as to whether or not the character string matches the character chain information of the 21st data to the 24th data. And a comparing means for judging the presence or absence of continuation of the chain.

[0041]

Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment 1) FIG. 1 (a) is a schematic diagram of a case where a character string collating apparatus is constructed by calculation using a recording medium according to the present invention, and FIG. 1 (b) is a character string according to the present invention. FIG. 2 is a block diagram of the collation apparatus, and FIG. 2 shows the concept of the character string collation method according to the first method of the present invention and the storage format of a recording medium that stores full-text search data.

In FIG. 2A, reference numeral 201 denotes a character string "iro a ha ni" inputted at the time of registration; 202, a two-character chain "iro" to be registered first; “Roaha” and 204 are the two-character chain “Hani” next to 203. Here, "a" indicates a special character inserted into the character string to separate the meaning.

In FIG. 2 (c), reference numeral 211 denotes a search character string "iro a ha ni" at the time of search, 212 denotes a two-character chain "iro" to be searched first, and 213 denotes a three-character chain "ro" next to 212. a is "," 214 is a two-character sequence "Hani" following 213.

In FIG. 2B, the two-letter chain 202 determines the number of appearances n1 and n2 of “i” and “ro” by the three-letter chain 2
03 stores the number of appearances n2 and n3 of “ro” and “ha”, and the two-character chain 204 stores the number of appearances n3 and n4 of “ha” and “ni”. Two-character chains 202 and 204 and three-character chains 203
Is stored in a different area and distinguishes between a two-character chain and a three-character chain.

With respect to the input of FIG. 2 (c), in the collation method according to the first method of the present invention, a two-character chain 202 corresponding to the "color" of the two-character chain 212 is stored as a two-character chain. Detected from the area, the number of appearances of "ro" at this time n2
And a three-letter chain 203 “roha” corresponding to “ro aha” in a three-letter chain 213 next to 212 is detected from the area in which the three-letter chain is stored, and the number of appearances n2 of “ro” at this time is detected. It is determined whether or not matches. If they match, then the number of appearances n3 of “ha” detected in 203 and the two-character chain 204 corresponding to “hani” in the two-character chain following 213 are detected from the area storing the two-character chain. It is determined whether or not the number of appearances of “ha” at the time matches. If they match, it is determined that the character string 211 matches 201. As described above, the character strings are collated.

FIG. 1B shows the configuration of a character string collating apparatus according to an embodiment of the first method of the present invention.

In FIG. 1B, reference numeral 101 denotes a two-character chain detector for detecting two-character chains 202 and 204 to be registered from a character string 201 to be registered, and 102 denotes a character string 2 to be registered.
A three-character chain detector 103 detects a three-character chain 203 registered from 01, a two-character chain memory 103 for storing two-character chains 202 and 204 and the number of appearances of those characters, and
04 is a three-character chain memory for storing the three-character chain 203 and the number of appearances of the first and last characters of the chain; 111 is a two-character chain 212 to search from the character string 211 to be searched;
A two-character chain detector 214 for detecting 214; a three-character chain detector 112 for detecting a three-character chain 213 to be searched from the character string 211 to be searched; 113 a two-character chain 212 detected by the two-character chain detector 111; 214 in the two-character chain memory 103 or the three-character chain detector 11
The three-character chain 213 detected from the second character chain is detected by the three-character chain memory 104, and whether the number of appearances of the character before each detected character chain matches the number of occurrences of the character after the character chain detected immediately before is detected. A comparator 114 determines whether or not all two- or three-character chains detected by the two-character chain detector 111 and the three-character chain detector 112 match with each other. Is a control unit that determines

In the above-described configuration, in the schematic diagram of FIG. Corresponding to

In the character string collating device configured as described above, the two-character chain memory 103 stores 202, FIG.
The two-character chain 204 is stored in the three-character chain memory as shown in FIG.
2 is stored, and an operation in the case where "iroa wa ni" in FIG. 2C is input as a search character string will be described.

When the search character string "color a han" is input, the two-character chain detection means converts the two character chain "color" and "hani" which do not include "a" specified as a special character in advance. Detected and output to the comparator 113. The three-letter chain detector detects the three-letter chain “roaha” in which “a” is inserted, centering on “a” specified as a special character in advance, and outputs it to the comparator 113.

At this time, the output to the comparator may be in the order of the sequence "color", "color a", "color", or together with the character chain information, "color", "color a", "color". May be output at the same time.

The comparator 113 discriminates between the output from the two-character chain detector and the output from the three-character chain detector. And a chain of “roha” corresponding to “ro a” is detected, and the chain is determined based on the number of appearances.

The comparator distinguishes between a two-character chain and a three-character chain, and detects them from different chain memories, thereby performing a search by distinguishing between "iro a ha ni" and "iro ha ni" as character strings to be searched. Becomes possible.

As described above, according to the present embodiment, it is possible to avoid an increase in the chain memory due to special characters without being limited by the number of appearances of the specific character “a” specified in advance. At the same time, it is possible to efficiently perform a chain extraction process based on the coincidence of the number of appearances.

In this embodiment, the special character is "a".
However, by replacing the special character sequence “a, a..., A” with “a”, a character that distinguishes whether a special character is inserted or not is not limited by the number of appearances of the special character. String matching by chaining can be performed.

That is, “Iro (one or more special characters)
And "irohani" can be searched using different search characters.

In this embodiment, different chain memories are provided to distinguish between a two-character chain and a three-character chain (insertion of special characters). However, the same memory is used to distinguish between a two-character chain and a three-character chain. It is also possible to provide an identifier and store it for a three-character chain, for example, as shown in FIG. (Embodiment 2) FIG. 3 is a conceptual diagram showing a configuration of a character string collating apparatus according to a second embodiment of the present invention. FIG. 4 is a conceptual diagram of a second method of character string collating according to the present embodiment. And a storage format of a recording medium storing full-text search data.

In FIG. 4A, reference numeral 401 denotes a character string "iro a ha ni" inputted at the time of registration, and 402 denotes a specific special character "a" for the character string 401 by the subsequent character "ha". The character string "iro a1 ha ni" changed to "a1" determined in 403, 403 is a two-character chain "iro" to be registered first, and 404 is a two-character chain "ro a1" next to 403.
05 is the next two-letter chain "a1" after 404, 406 is 4
It is the two-letter chain "Hani" following 05.

Here, "a" is a special character inserted into a character string to separate meanings and the like, and "a1" represents a specific symbol or code not to be searched.

In FIG. 4C, reference numeral 411 denotes a search character string “Iro-a-ha-ni” at the time of search, and reference numeral 412 denotes a specific special character “a” for the character string 411 by the subsequent character “Ha”. Character string changed to "a1" decided "color a1 hanni", 4
13 is a two-letter chain "iro" to be searched first, 414 is a two-letter chain following the 413 "ro a1", 415 is a two-letter chain following the 414 "a1", and 416 is a two-letter character following the 415 The chain is "Hani".

In FIG. 4B, the two-letter chain 403 is the number of appearances n1 and n2 of the search target of “i” and “ro”, and the two-letter chain 404 is “ro” and “a1”.
The two-letter chain 405 stores the occurrence counts n3 and n4 of “a1” and “ha”, and the two-letter chain 406 stores the occurrence counts n4 and n5 of “ha” and “ni”.

At this time, in the collation method according to the second method of the present invention, the two-character chain 403 corresponding to the “iro” of the two-character chain 413 is detected, and the number of appearances n2 of “ro” at this time is detected.
, A two-character chain 404 corresponding to “ro a1” in the two-character chain 414 following 413 is detected, and it is determined whether or not the number of appearances n2 of “ro” at this time matches. If they match, then the number of appearances n3 of “a1” detected in 404 and the two-character sequence 405 corresponding to “a1” in the two-character sequence following 414
Is detected, and it is determined whether or not the number of appearances of “a1” at this time matches. If there is a match, then the number of appearances n4 of "ha" detected at 405 and the two-character chain 406 corresponding to "hani" of the next two-character chain of 415 are detected, and the number of occurrences of "ha" at this time It is determined whether or not matches. If they match, it is determined that the character string 411 matches 401. As described above, the character strings are collated.

FIG. 3 shows the configuration of a character string collating apparatus according to an embodiment of the second method of the present invention.

In FIG. 3, reference numeral 301 denotes a character string 4 to be registered.
01 is a character string converter that changes a specific special character “a” to a character string 402 that is uniquely determined by the subsequent character “ha”, and 302 is a two-character chain 403 registered from the character string 402 , 404, 405, 406, and 303 is a two-character chain 403, 40.
4, 405, 406, and a two-character chain memory for storing the number of appearances of those characters.
1 is a character string converter that changes a specific special character “a” to a character string 412 that is uniquely determined by the subsequent character “ha”, and 305 is a two-character chain 413 to search in the character string 412. , 414, 415, and 416, a two-character chain detector 306, and a two-character chain detector 305
Two character chains 413, 414, 415, 4
307, a comparator 307 that detects 16 in the two-character chain memory 303 and determines whether the number of appearances of the character before the detected two-character chain matches the number of appearances of the character after the two-character chain detected immediately before. Is a control unit for determining by the comparator 306 all the two-character chains detected by the two-character chain detector 305, and determining whether the character strings match.

The operation of the thus constructed character string collating apparatus will be described. When the registered character string is input, the character string conversion unit 301 converts the special character “a” specified in advance into a symbol or code that is not a predetermined search target by a subsequent character, that is, a symbol other than the search character string, Convert to code and output.

In the character string conversion means, as shown in FIG. 4D, the correspondence to the character after the special symbol and the conversion to which symbol is stored. This correspondence is 421, 422
, Or may correspond to a group of characters, such as 423.

In the converted character string, a two-character chain and its appearance and recovery are detected by a two-character chain detector as in the first embodiment, and are stored in a two-character chain memory.

On the other hand, when the search character string is given, the character string converter 304 converts the special character into a symbol or code other than the search character string according to the same correspondence as that used in the character string conversion 301, and converts the special character into two characters. Output to the chain detector. The two-character chain detector detects the two-character chain and outputs the result to the comparator 306.

The comparator 306 detects the coincidence of the character chains according to the contents of the two-character chain memory according to the same procedure as in the first embodiment. However, in the second embodiment, it is not necessary for the comparator to distinguish between a two-character chain and a three-character chain as in the first embodiment.

As described above, according to the present embodiment, it is possible to avoid an increase in the chain memory of the special character “a” having a high appearance frequency, and to replace the same special character with a plurality of different characters in accordance with the subsequent characters. By converting to characters, candidates for checking the coincidence of the number of appearances for extracting a chain are dispersed into a plurality of candidates, so that the processing time is reduced.

In this embodiment, the special character "a"
Is determined based on the character after that, but it is clear that the same effect can be obtained even when the conversion destination is determined based on the character before the special character.

The schematic diagram of the case where the present invention is implemented as a computer is the same as that of FIG. 1A.

(Embodiment 3) FIG. 5 is a block diagram showing a configuration of a character string collating apparatus according to a third embodiment of the present invention. FIGS. 2 shows the concept of the method and the storage format of a recording medium that stores full-text search data.

In FIG. 6, reference numeral 601 denotes a character string “iroaha ni” inputted at the time of registration, 602 denotes a specific special character “a” for the character string 601 and the character “ro” preceding it is “ro”.
And "ro", which is uniquely determined by "ro", and the subsequent character "ha", which is uniquely determined by "ha", "ha", which is composed of "ha" and "ha""603" is a two-character chain "iro" that is registered first, 604 is a two-character chain "roro" next to 603, and 605 is a character sequence "604". The next two-letter chain "ro'ha '", 606 is the next two-letter chain after 605 "ha'ha", and 607 is the two-letter chain next to 606 "hani"
It is.

Here, "a" is a special character inserted into a character string for separating meanings, etc.
Represents a specific symbol or code not to be searched.

In FIG. 8, reference numeral 611 denotes a search character string “iro a ha ni” at the time of search, and reference numeral 612 denotes a specific special character “a” for the character string 611 and the character “ro” preceding it is “ro”. And "ro", which is uniquely determined by "ro", and the subsequent character "ha", which is uniquely determined by "ha", "ha", which is composed of "ha" and "ha""613 is the two-character chain" iro "to be searched first, and 614 is the next two characters after 613.
The character chain "Roro '", 615 is the two-character chain next to 614 "Roh'", and 616 is the two-character chain next to 615 "Hana"
"Ha", 617 is a two-character chain "hani" next to 616.

In FIG. 7, the two-character chain 603 is “I”.
And the number of appearances n1 and n2 of "ro", the two-character chain 604 is the number of appearances n2 and n3 of "ro" and "ro '", and the two-character chain 605 is the number of appearances of "ro'" and "ha '" n3, n4,
The two-character chain 606 indicates the number of appearances n of “ha” and “ha”.
The four-character chain 607 stores the number of appearances n5 and n6 of “ha” and “ni”. At this time, in the collation method according to the third method of the present invention, a two-character chain 603 corresponding to “iro” of the two-character chain 613 is detected, and the number of appearances n2 of “ro” at this time and the next two The two-character chain 604 corresponding to “Roro ′” in the character chain 614 is detected, and it is determined whether or not the number of appearances n2 of “Roro” at this time matches. If they match, then the number of appearances n3 of “ro '” detected in 604
, A two-character chain 605 corresponding to “ro'wa” in the two-character chain following 614 is detected, and it is determined whether the number of appearances of “ro '” at this time matches. If they match, then 60
The number of appearances n4 of “wa '” detected in 5 and the next 2 in 615
The two-character chain 606 corresponding to the character chain “ha′ha” is detected, and it is determined whether or not the number of appearances of “ha ′” at this time matches. If there is a match, then the number of occurrences n5 of “ha” detected in 606 and the two-character chain 607 corresponding to “hani” of the two-character chain following 616 are detected, and the number of occurrences of “ha” at this time It is determined whether or not matches. If they match, it is determined that the character string 611 matches 601. From the above,
String matching is performed.

FIG. 5 shows a configuration of a character string collating apparatus according to an embodiment of the third method of the present invention.

In FIG. 5, reference numeral 501 denotes a character string 6 to be registered.
01, the special character "a" is replaced by "ro" consisting of "ro", which is uniquely determined by "ro" and "ro", and the subsequent character "ro""Is a character string converter that is uniquely determined by" ha "and is changed to a character string 602 changed to"ha'ha"consistingof" ha '"and" ha ". 502 is a two-character chain 603 registered from the character string 602. , 604, 605, 606, 607
503 is a two-character chain that detects
3, 604, 605, 606, 607 and the two-character chain memory for storing the number of appearances of those characters. 504 is a character string 611 to be searched. "Ro" consisting of "ro" uniquely determined by "" and "ro", and the following character "ha" uniquely consisting of "ha" and "ha" consisting of "ha" and "ha" A character string converter for changing to a character string 612 changed to "ha". Reference numeral 505 denotes a two-character chain 613, 614, 615, 616, 6 to be searched in the character string 612.
17 is a two-character chain detector 506, and the two-character chain 613 and 61 detected by the two-character chain detector 505.
4, 615, 616, and 617 are stored in the two-character chain memory 503.
The comparator 507 determines whether or not the number of occurrences of the character before the detected two-character chain matches the number of occurrences of the character after the two-character chain detected immediately before. The controller 506 determines all the two-character chains detected from, and determines whether the character strings match.

As described above, according to the present embodiment, it is possible to perform character string collation using character chains without being limited by the number of appearances of the special character “a”.

That is, according to the embodiment, the special character “a”
Is converted into separate characters by the characters before and after it, and the number of appearances of the converted characters is recorded. Therefore, compared to the second embodiment, the two-character chain file is further finely dispersed, so that the frequency of use is high. It is possible to avoid an increase in the chain memory of the special character “a” having a high appearance frequency of the special character,
At the same time, the efficiency of the chain extraction process can be improved.

The schematic diagram in the case of being implemented as a computer is the same as FIG. 1A, and in this case, the two-character chain memory 503 corresponds to the external recording device 40. (Embodiment 4) FIG. 10 shows the fourth embodiment of the character string collation according to the present invention.
Shows the concept of the method. In FIG. 10 (a), 1
001 is a character string “Iro a Hani” entered at the time of registration, 1
002 is a two-character chain “Iro” that is registered first, 100
3 is the first two-character chain “loha” generated from the three-character string “roaha” sandwiching the special character “a” following 1002,
1004 is a two-character chain “roa” including a special character generated next to 1003, 1005 is a two-character chain “a wa” including a special character generated next to 1004, and 1006 is 1005
Is the next two-letter chain "Hani". In FIG. 10 (c), reference numeral 1011 denotes a search character string at the time of search,
012 is a two-character chain “iro” to be searched first, 101
3 is the first two-character chain “loha” generated from the three-character string “roaha” sandwiching the special character “a” following 1012,
1014 is a two-character chain “roa” including a special character generated next to 1013, 1015 is a two-character chain “a wa” including a special character generated next to 1014, and 1016 is 1015
Is the next two-letter chain "Hani".

In FIG. 10 (b), a two-character chain 100
2 is the number of appearances n1 and n2 of “i” and “ro”, the two-character chain 1003 is the number of appearances n2 and n3 of “ro” and “ha”,
The two-character chain 1004 includes the number of appearances n2 of “ro” and “a”
Is a constant value n, a two-character chain 1005 is a constant value n and the number of appearances n3 of “ha” for “a”, a two-character chain 1006 is a number of appearances n3 of “ha” and “ni”, Remember n4.

At this time, in the collation method according to the fourth method of the present invention, a two-character chain 1002 corresponding to “iro” of the two-character chain 1012 is detected, and the number of appearances of “ro” at this time is detected.
n2 and a three-character string "roa" sandwiching "a" next to 1012
Is a two-character chain 1013 out of the two-character chain generated from
Then, a two-character chain 1003 corresponding to "roha" is detected, and the number of appearances n2 and n3 of "ro" and "ha" at this time are detected. It is determined whether or not the number of appearances of “ro” in the character chains 1002 and 1003 matches with n2. If they match, a character chain 1004 corresponding to “a” in the character chain 1014 is next.
Is detected, and it is determined whether the number of appearances of “ro” is n2. Next, the character chain 1 corresponding to “a” in the character chain 1015
005 is detected, and it is determined whether the number of appearances of “ha” is n3 and coincides with the number of appearances n3 of “ro” detected in 1003. If there is a match, then the number of appearances n3 of “ha” detected in 1005,
A two-character chain 1006 corresponding to “Hani” in the next two-character chain after 1015 is detected, and it is determined whether or not the number of appearances of “Hana” at this time matches. If they match, the character string 1011
Is determined to match 1001. As described above, the character strings are collated.

FIG. 9 shows the configuration of a character string collating apparatus according to an embodiment of the fourth method of the present invention.

In FIG. 9, reference numeral 901 denotes a special character detector for detecting a specific special character “a” in the character string 1001 to be registered, and reference numeral 902 denotes a registration 2 generated when there is no special character from the character string 1001. Character chains 1002, 100
6, a two-character chain detector 903 detects the character string 100
A special character chain detector for detecting a two-character chain 1003, 1004, 1005 generated from a three-character string "roaha" sandwiching the special character "a" from 1; 904, a two-character chain detector 902 and a special character chain detection 2 detected by the detector 903
Character chains 1002, 1003, 1004, 1005, 1
006 and each chained character, a special character is a fixed value, and the other characters are a two-character chain memory that stores the number of appearances. 911 is a special character that detects a specific special character “a” in the character string 1011 to be searched. Character detector, 912
Is a two-character chain detector that detects two-character chains 1012 and 1016 to be registered, which are generated when there is no special character from the character string 1011; A special character chain detector that detects the two-character chain 1013, 1014, and 1015 generated from "ha", and 914 detects the two-character chain 1012 and 1016 detected by the two-character chain detector 912 in the two-character chain memory 904, It determines whether or not the number of appearances of the character before the detected two-character chain matches the number of occurrences of the character after the two-character chain detected immediately before.
13 two-character chain 1013, 1014, 1
015 is detected by the two-character chain memory 904, and a comparator that determines whether the number of appearances of “ro” and “ha” other than special characters matches in the detected character chain, and 915 is a two-character chain detector 912 and a special character The controller 914 determines all the two-character chains detected by the character chain detector 913 with the comparator 914 and determines whether the character strings match.

Therefore, in this method, a specific special character “a” can be linked with the preceding and following characters irrespective of its appearance frequency. String matching by chaining can be performed.

(Embodiment 5) FIG. 12 shows a fifth embodiment of the present invention.
3 shows the concept of the method of character string matching. Twelfth
In FIG. 12A, reference numeral 1201 denotes a character string “Iro a Hari” input at the time of registration; 1202 denotes a two-character chain “Iro” which is registered first; , 1204 is a two-character chain “Hani” next to 1203. In FIG. 12 (c), reference numeral 1211 denotes a search character string "iro a ha ni" at the time of search, reference numeral 1212 denotes a two-character chain "iro" to be searched first, and reference numeral 1213 denotes a character string next to 1212.
The character chain “roaha” and 1214 are the next two-character chain “hani” after 1213.

In FIG. 12B, the two-character chain 120
2 is the number of appearances n1 and n2 of “i” and “ro”, and the three-character chain 1203 is the number of appearances n2 of “ro” and 0 of the number of “a”.
And the combination of the number of times “a” 0 and the number of appearances n3 of “ha”, n2, 0 and 0, n3. Remember.

At this time, in the collation method according to the fifth method of the present invention, a two-character chain 1202 corresponding to “iro” of the two-character chain 1212 is detected, and the number of appearances of “ro” at this time is detected.
n2 and "roaha" in the three-character chain 1213 following 1212
Is detected, and it is determined whether or not the number of appearances n2 of “ro” at this time matches. If they match, a value 0 corresponding to "a" in the three-character chain is detected. Next, the value 0 of “a” before “ha” is detected, and 12
The number of appearances n3 of “ha” detected in 03 and the two-character chain 1204 corresponding to “hani” of the two-character chain next to 1213 are detected, and whether or not the number of appearances of “ha” at this time matches Judge. If they match, it is determined that the character string 1211 matches 1201. As described above, the character strings are collated.

FIG. 11 shows the configuration of a character string collating apparatus in one embodiment of the fifth method of the present invention.

In FIG. 11, reference numeral 1101 denotes a two-character chain 1202, 120 to be registered from a character string 1201 to be registered.
4, a two-character chain detector 1102 detects a three-character chain 1203 to be registered from a character string 1201 to be registered, and a two-character chain detector 1103 detects a two-character chain 1203.
2, a two-character chain memory for storing 1204 and the number of appearances of those characters; and 1104, a three-character chain 1203 for storing the number of appearances of the first and last characters of the chain.
The character chain memory 1111 detects two-character chains 1212 and 1214 to be searched from the character string 1211 to be searched.
A character chain detector 1112 is a three-character chain detector that detects a three-character chain 1213 to be searched from a character string 1211 to be searched, and 1113 is a two-character chain of two-character chains 1212 and 1214 detected by the two-character chain detector 1111. Detected in memory 1103 or three-character chain detector 111
The three-character chain 1213 detected from the second character chain is detected by the three-character chain memory 1104, and whether the number of appearances of the character before each detected character chain matches the number of occurrences of the character after the character chain detected immediately before is detected. Comparator for determining whether or not 1114
Are the two-character chain detector 1111 and the three-character chain detector 1
The controller 1113 determines whether the two or three character chains detected from the sequence 112 match with each other by the comparator 1113 and determines whether the character strings match.

Therefore, at this time, it is possible to perform character string collation by character chains without being limited by the number of appearances of the specific special character “a”.

(Embodiment 6) FIG. 14 shows a fifth embodiment of the present invention.
3 shows the concept of the method of character string matching. 14th
In FIG. 13A, reference numeral 1401 denotes a character string “Iro a Hari” input at the time of registration; 1402 denotes a two-character chain “Iro” which is registered first; The special character "a" is converted to the next character "ha" by the second character of the character chain "roaha". The two-character chain "roha" of the first character and the second character of the three-character chain 1403 is converted, and 1405 is the two-character chain "haha" of the second character and the third character of the three-character chain 1403. It is a two character chain "Hani". In FIG. 14 (c),
1411 is a search character string at the time of search “iro a hani”, 141
2 is a two-character chain “iro” to be searched first, and 1413 is a three-character chain “ro a” in which the special character next to 1412 is inserted.
The special character “a” is the second character of “ha” and the next third character is “ha”
1414 converts the special character "a" to the next character and converts the special character "a" into the next character and converts the special character "a" to the next character.
The two-character chain “Hana” consisting of the second and third characters in the character chain 1413 is shown, and 1416 is the two-character chain “Hani” next to 1415.

In FIG. 14B, a two-character chain 140
2 is the number of appearances n1 and n2 of “i” and “ro”, and the two-character chain 1404 is the third number of appearances n2 and 1403 of “ro”.
The combination of the number of appearances n3 of the character "ha" is n2, n3, and the two-character chain 1405 is the combination of the number of appearances n3 of the third character "ha" of two 1403, n3, n3 is the combination of two characters 1204.
Stores the number of appearances n3 and n4 of “ha” and “ni”.

At this time, in the collation method according to the sixth method of the present invention, a two-character chain 1402 corresponding to “iro” of the two-character chain 1412 is detected, and the number of appearances of “ro” at this time is detected.
n2 and a two-character chain 140 corresponding to “roha” of the first two-character chain 1414 of the three-character chain 1413 following the 1412
4 is detected, and it is determined whether or not the number of appearances n2 of “ro” at this time matches. If they match, then the next two in the three-character chain
Two-character chain 140 corresponding to character chain 1415 "haha"
5 is detected, and the number of appearances n3 of "ha" at this time matches, and the number of occurrences of the first character "ha" and the number of appearances of the second character "ha" in the chain of 1405 match at n3. To detect. Next, a two-character chain 1406 corresponding to "Hani" in the two-character chain 1416 is detected, and the number of appearances n3 of "H" detected in 1405 matches the number of occurrences of "Hana" in the two-character chain 1406? Determine whether or not. If they match, the character string 1411 is 1
It is determined that the number matches 401. As described above, the character strings are collated.

FIG. 13 shows the structure of a character string collating apparatus according to an embodiment of the sixth method of the present invention.

In FIG. 13, reference numeral 1301 denotes a two-character chain 1402, 140 to be registered from a character string 1401 to be registered.
6, a two-character chain detector 1302 that detects the number of appearances of each character,
Three-character chain detector 130 for detecting character chain 1403
3 is a special two-character chain generator for changing the special character inserted from the three-character chain 1403 to the next character and detecting two two-character chains 1404 and 1405 and the number of appearances of each character. 1402, 1404, 14
05, 1406 and a two-character chain memory for storing the number of appearances of those characters.
2 is a two-character chain detector that detects two-character chains 1412 and 1416 to be searched from 1, and 1312 is a character string 141 to be searched.
A three-character chain detector 1313 for detecting a three-character chain 1413 to be searched from 1 is used to change the special character inserted from the three-character chain 1413 to the next character, and 1414 and 141
5 is a special two-character chain generator for detecting the two two-character chain and the number of appearances of each character.
The two-character chains 1412 and 1416 detected from the 311 are detected by the two-character chain memory 1304 or the special 2
Two-character chain 14 generated by character chain generator 1313
14, 1415 are detected by the two-character chain memory 1304,
It is determined whether or not the number of occurrences of the character before the detected character chain matches the number of occurrences of the character after the character chain detected immediately before. In the case of the special two-character chain 1415, the first character and the second character are determined. A comparator 1315 determines that the appearance counts of the characters match each other. A comparator 1314 determines whether the two or three character chains detected by the two-character chain detector 1311 and the three-character chain detector 1312 match. It is a control unit that determines and determines whether the character strings match.

Therefore, at this time, it is possible to perform character string collation using character chains without being limited by the number of appearances of the specific special character “a”.

(Embodiment 7) FIG. 15 shows a seventh embodiment of the present invention.
3 shows the concept of the method of character string matching. Fifteenth
Referring to FIG. 15A, reference numeral 1501 denotes a character string "color a wa ni a color a hat a" input at the time of registration; "a", 1502
Is the next two-letter chain "aha" including 1503, and similarly generates a two-letter chain from 1505 to 1512.

In FIG. 15B, a two-character chain 15
02 indicates the number of appearances of “i” and “ro” as n1 and n2, and the two-character chain 1505 indicates the number of occurrences of “ha” and “ni” n3 and n4
, The two-character chain 1508 is the number of appearances n1 +
1, n2 + 1, the two-character chain 1511 is "ha" and "to"
Are stored n3 + 1 and n5. For example, FIG. 15 (e) shows a set of stored appearance counts of the two-character chain “hato”.

Next, the maximum value of the number of appearances of the special character "a" is designated as 2 in advance, and if the remainder obtained by dividing the number of occurrences of the special character by the maximum value is 0, the maximum value is set. To be specified. In this case, the number of appearances of the special character is one or two. In FIG. 15 (b), the two-character chain 15
03 is the first occurrence of the special character “a”, so the number of appearances is 1, the number of appearances of the special character “a” in the character chain 1504 is also the same, and the special character “a” in the two-character chain 1506 is 1
Since it is the second appearance, the number of appearances is 2, character chain 1507
The number of appearances of the special character “a” is also 2. on the other hand,
The special character “a” in the two-character chain 1509 has the appearance frequency 1 of 2
Since this is the second appearance, the number of appearances is 1, and the character chain 1510
The number of appearances of the special character "a" is also 1, and the character chain 15
Since the twelve special characters “a” appear for the second time, the number of appearances is two.

Next, the two-character chain including the special character is stored by sorting the second character for each character type. In FIG. 15 (c), the set of the character chain of the two-character chain "roa" is the set n2,1 of the two-character chain 1503, and the two-character chain n2 + 1.
It is composed of 1. On the other hand, in FIG. 15 (d), a character chain set of a two-character chain "a *", * is composed of the character types "i" and "ha" that appear, and a two-character chain 1504;
1507 and 1510 are sorted for each character type. Here, the sorting for each character type is in the order of the character code, and when the number of appearances matches, the order in which the characters appear in the registered character string. As a result of the sorting, the character chain is as shown in FIG. 15 (c).
It is stored as shown in FIG.

At this time, in the collation method according to the sixteenth method of the present invention, a description will be given by taking as an example the search character string "roa-hato" shown in FIG. 15 (f).

First, the continuity of the character chains “a” and “a” are collated. When the matching is started, “a” and “a”
Is reset to zero. Character chain 15
15 (c) 1503 is first detected for the two-character chain corresponding to the 13 "roa", and the "roa" overlap shown in FIG. The counter stores the number of occurrences 0 of the number of appearances 1 in the counter. Next, regarding the two-character chain corresponding to “a wa” of the character chain
5 In FIG. 5D, the number of appearances is detected in order from the first character chain of “a wa”, and the number of times of duplication 0 is stored in the “a wa” duplication counter. The collation is the number of appearances of the second character in the two character chains 1503 and the number of appearances of the first character in 1504,
It is checked whether or not the number of repetitions of “a” and “a” match, and if they do, the character chain “hato” is further collated. Here, since the number of appearances of the second character in the character chain 1504 is different from the number of appearances of the first character in the character chain 1510, the next character chain is collated. Character chain 1
The number of occurrences of special characters in the character chain 503 and the character chain 1509 is checked for overlap.
6 is increased by one. Thereby, the duplication counter 1517 of “a wa” corresponding to the character chain 1510 is increased by one. Subsequently, in FIG. 15 (d), the duplicate counter 15 of "a"
Since 16 is 1, a character chain 1510 in which only one first character is duplicated is detected. The character chains 1509, 1510, and 1511 are finally detected as a continuous character string from the collation of the character sequence "hato". At this time, the duplicate counter value 1 (151) of “a” and “a” in FIG.
6, 1517) are stored.

As described above, the character strings are collated. Note that the maximum value of the special character can be arbitrarily specified. In the present embodiment, the number of occurrences of the special character is a remainder obtained by dividing the number of occurrences of the special character by the maximum value of the number of occurrences specified in advance. If the remainder is 0, the maximum value is set. Although the case is given, the number of appearances is not more than the maximum value and it is only necessary to ignore duplication, so it is sufficient that the number of occurrences is less than the maximum value of the number of appearances. There are odd numbers, odd numbers in descending order, and so on. For example,
Assuming that the maximum value is 10, the number of appearances of special characters is 3, 5,
The repetition of 7, 8, 6, 4, 2 may be used as the number of appearances.

FIG. 16 shows the structure of a character string collating apparatus according to an embodiment of the seventh method of the present invention.

In FIG. 16, reference numeral 1601 denotes a special character detector for detecting a specific special character “a” in a registered character string 1501, and 1602 denotes a character chain of a character chain when there is no special character from the character string 1501. And a two-character chain detector that calculates the number of appearances and stores 1502, 1505, 1508, and 1511 in the two-character chain memory 1606.
03 is obtained from the two-character chain memory 1606 to determine the number of appearances of the first character or the second character of a character type that does not include special characters in a two-character chain including special characters, and further reduces the number of occurrences of special characters to a maximum value or less. The number of occurrences of the special character is calculated and stored in the occurrence overlap memory 1604, and the value of the number of appearances of the next special character is calculated from the occurrence overlap memory 1604.
3, 1504, 1506, 1507, 159, 151
Special character chain detector to determine 0, 1512, 1605
Is the first special character from the special character chain detector 1605.
A special character chain sorter that sometimes sorts by character type of the second character as a character and stores the sorted result (FIG. 15 (d)) in a two-character chain memory 1606. Reference numeral 1607 denotes a search character string (FIG. 15). (F)) a special character detector for detecting the special character “a” from the search character string, 1608 is a two-character chain detector for generating a two-character chain from the search character string when there is no special character, and 1609 is a A two-character chain detector that generates a two-character chain including special characters,
The two-character chain memory 1606 stores the character chains corresponding to the character chains 1513, 1514, and 1515 detected by the character chain detector 1608 and the special character chain detector 169 and the number of appearances.
From the two character chains 1513 and 1514, the appearance duplication counter memory 1612 is set to 0, and
The number of duplications between the character chain 1503 and the second character is 0, the number of duplications between the two-character chain 1504 and the first character is calculated as 0, and then the number of duplications between the two-character chain 1509 and the second character is 1, and the two-character chain 1510 and the Calculate the number of duplications of 1 character 1 and 2 character chains 1
For 515, a comparator for calculating 1511, 1611
Is a controller for judging a match between character strings from the results calculated by the two-character chain detector 1608 and the comparator 1610.

Therefore, at this time, it is possible to perform character string collation by character chain without being limited by the number of appearances of the specific special character “a”.

(Embodiment 8) FIG. 17 shows an eighth embodiment of the present invention.
2 shows the concept of the method of registration and the method of character string collation. First, a registration method will be described.

In FIG. 17 (d), reference numeral 1708 denotes a character string "ai aai aai aaiai" input at the time of registration, 1709 denotes a two-character chain "ai" registered first, and 1710 denotes a special character following 1709. The two-letter chain "a", 1711 including the character "a" is converted to the next two-letter chain "a".
And the same applies to 1712-1720.
Generate a character chain. From this two-character chain, character chain data, which is a set in which a document number, the number of appearances of the first character and the second character, or a numerical value is stored, is generated.

FIGS. 17 (a) to 17 (c) show different character chain data according to the type of characters constituting the two-character chain. FIG. 17A shows character chain data that does not include special characters, and has the same size for storing the number of appearances of the first character and the number of appearances of the second character. On the other hand, FIGS. 17 (b) and (c)
Is character chain data including special characters, and the size of the number of appearances for special characters is larger than the area for characters that are not special characters. A designated value (0 in this embodiment) is stored in an area for a character that is not a special character.

In FIG. 17 (e), the registered character string 17
08 to create character chain data. Here, the number of appearances of “A” is n1, and the number of appearances of “I” is n2. The two-character chain 1709 is character chain data of “Ai”, and “A”
Are n1 and "i" n2, but the character chain data of the two-character chain "a" 1710 following the two-character chain "ai" is composed of FIG. In the character chain data of "", the value corresponding to "i" is 0, and the value corresponding to "a" is 1 the number of special character appearances. Therefore, from the continuity of the character chain data, the character chain data of “A” is 172 characters.
The portion corresponding to the second character becomes 0, such as 1. Hereinafter, similarly, character chain data such as 1722 to 1732 can be constituted.

The character chain data generated in FIG. 17 (e) is stored separately for each combination of two character chains appearing as shown in FIGS. 17 (f) to 17 (i).

The above-described method of generating character chain data is based on the first method.
This is realized by the flow of FIG. 8 and the configuration of the character string collation device of FIG. In FIG. 20, 2001 is a two-character chain detector that creates a two-character chain and a document number from a registered character string, and 2002 calculates the number of appearances or value for each character type of the two-character chain from the two-character chain, Furthermore, in the case of a two-character chain following a two-character chain including special characters,
An inquiry is made to the 2003 special character chain detector as to whether it is necessary to replace the value of the number of appearances with a value already calculated, and an appearance number calculator for calculating the number of appearances again from the result. A two-character chain memory 20 for storing a set of occurrence counts as character chain data
04. In FIG. 18, the character chain detector 2
001 reads the registered character string data (step 180
1), a document number is set (step 1802), a registered character string is read up to the maximum number of documents, a document number is assigned (step 1803), and a two-character chain (Ak, Ak + 1) (Ak, A
Create a set of (k + 1 is the k, k + 1th character type) (Step 1)
804). Subsequently, the appearance number detector 2002 checks the presence or absence of a special character string in the two-character chain (step 1805). When the special character is included, the appearance number N (Ak) or N
(Ak + 1) is counted, and character chain data Sk is created (steps 1806 and 1808). If no special character is included, the number of appearances is counted to create character chain data Sk (step 1810). Next, the special character chain detector 20
03 is such that the number of appearances or value corresponding to the second character of Sk is equal to the number of appearances or value corresponding to the first character of Sk + 1 with respect to the character chain data Sk + 1 continuous with the character chain data Sk. The value is corrected so as to satisfy (steps 1807 and 180
8, 1811). The following is performed for all two-character chains and registered character strings (steps 1812 to 1814),
The generated character chain data is stored in the two-character continuous memory 2004.

Next, a method of character string collation will be described.
The search character string will be described using "Ia aa" in FIG. 17 (j) 1732 as an example. For the search character string, a two-character chain “a” 1733, a next two-character chain “a” 1734, and so on are created in the same manner from 1735 to 1736. This 2
Character chain data corresponding to the character chain is retrieved from the two-character chain memory 2004, and continuity is collated sequentially from 1733. The concept of the continuity check is shown in FIG. 17 (k). The character chain data corresponding to the two-character chain “Ia” 1733 is searched from the beginning in FIG. 17G that is the character chain data of “Ia”, and the character chain data 1722 is extracted. The character type of the character chain data 1722 is checked, and the special character “a” specified in advance is included in the second character.
The number of occurrences of “a” is stored in the special character appearance counter memory (second
(2007) in FIG. Next, the two-letter chain "Ia"
The character chain data corresponding to the two-character chain “a-a” following the character string “a-a” is searched from the beginning in FIG. It is checked whether the number of appearances of the first character in the character chain data of FIG. 17 (h) matches, and character chain data 1723 is obtained. As a result, character chain data 1722 and 172
3 is determined to be continuous.

Next, character chain data corresponding to the two-character chain "A" following the two-character chain "a" is searched from the beginning in FIG. It is checked whether the number of appearances of the second character in the data 1723 matches the number of appearances of the first character in the character chain data in FIG. 17F, and character chain data 1724 is obtained. Thus, the character chain data 1723 and 1724 are determined to be continuous.

Next, the character chain data corresponding to the two-character chain "Ia" following the two-character chain "A" is searched from the top in FIG. 17 (g) which is the character chain data of "Ia". It is checked whether the number of appearances of the second character in the chained data 1724 and the number of appearances of the first character in the character chained data in FIG. 17 (g) match. Here, the special character appearance counter 173 appears because the special character “a” appears again in the two-character chain “a”.
The value of 8 is increased by one (1738). From the top of FIG. 17 (g), when character chain data that matches the number of appearances of the second character of the character chain data 1724 and the value of the first character of the character chain data of “Ia” is checked, there is 1722. Since the number of occurrences of the special character is 2 from the special character appearance counter 1738, the next character chain data is searched for 1725. As a result, the character chain data 1724 and 1725 are determined to be continuous, and it is determined that a registered character string including the search character string exists.

The above-described method of character string collation is realized by the flow of FIG. 19 and the configuration of the character string collation apparatus of FIG. In FIG. 20, reference numeral 2005 denotes a two-character chain detector for creating a two-character chain from a character string to be searched;
Examines each character type constituting the two-character chain, and if the two-character chain does not include a special character, the character chain data corresponding to the continuous character chain detected by the character chain detector 2005 is detected. By comparing the number of appearances of the second character of the character chain data with the number of appearances of the first character of the character chain data of the character chain following the character chain, it is determined whether or not the character chain as a search character string is continuous. When the two-character chain includes a special character, the character chain data corresponding to the continuous character chain searched by the character chain detector 2005 is compared with the number of appearances of the character in the same manner as the comparing means. A comparator that stores the number of appearances of the special character string specified in the special character appearance counter memory 2007, and compares the number of occurrences other than continuous character chains based on the fact that the number of occurrences does not overlap. From result of the comparator 2007 is configured the presence or absence of continuity of the continuity of the character chain data determining controller. In FIG. 19, the two-character chain detector 2005 reads the search character string (step 1901), and the two-character chain
Ak, Ak + 1 are created (step 1902), and a two-character chain is set from the beginning (step 1903).
6 takes out a two-character chain from the two-character chain detector 2005 until the continuity can no longer be checked (step 19).
04), and two sets of two-character chains (Ak, Ak + 1) and (Ak + 1, Ak +
2) (Ak, Ak + 1, Ak + 2 is the k, K + 1, K + 2nd character type) Character chain data Sl (N (Ak), N (Ak + 1)), Sm (M (Ak), M (Ak +
1)) (Sl, Sm is the l, m-th character chain data, N (Ak), M (Ak + 1)
Extracts the number of appearances or values of the character types Ak and Ak + 1 from the beginning (step 1905) and checks whether a special character is included in the two-character chain (step 1906). If special characters are included, the number of occurrences of special characters N (Ak)
Or, store N (Ak + 1) as T, then character chain data Sl
Is checked whether the number of appearances N (Ak + 1) of the second character of the character string and the number of appearances M (Ak + 1) of the first character in the character chain data Sm match (step 1908). Is set to the next character chain data Sm + 1 (step 1910),
Move to step 1905. It is determined whether or not the number of appearances matches, whether the special character is character chain data including the special character, and matches the number of appearances T of the special character (step 190).
9) If the condition of step 1909 is not satisfied, the result of determining that the character chain data is continuous is returned to the controller 2008 (step 1911), and the continuous two-character chain collation is started (step 1912).

At this time, if the number of occurrences of the specific special character "a" is larger than that of other character types, it is possible to perform character string collation by character chain. In the eighth method of the present invention, when performing a search including a special character at the beginning such as “a AI” in the search character string, the first character string data of “AI” is used.
Since the number of appearances of a character is 0, it is first determined whether the number of appearances of the first character is 0 in the character chain data of "A" without referring to the character chain data of "a". Can shorten the collation processing.

(Embodiment 9) FIG. 22 shows a ninth embodiment of the present invention.
3 shows the concept of the method of character string matching. 22nd
Referring to FIG. 9A, reference numeral 2201 denotes a character string “Iro a Hari” input at the time of registration; 2202 denotes a two-character chain “Iro” to be registered first; 2203 denotes a special two-character chain following 2202; The special two-character chain “Iha”, which is a set of the first character and the character “H” next to the special character “a”, or the two-character chain “Hani” following the special character “a”
This is the first character of (2204). In FIG. 22 (c), reference numeral 2205 denotes a search character string “iro a han” at the time of search, followed by a two-character chain 2206 “iro”, and the first character of 2206 and the character “ha” immediately after the special character. Character chain 22
07 "Iha" and the two-character chain "Hani" after the special character "a".

In FIG. 22B, a two-character chain 22
02 is the number of appearances n1 and n2 of “i” and “ro”, and the two-character chain 2203 is the number of appearances n1 and n3 of “i” and “ha”.
And the two-character chain 2204 stores the number of appearances n3 and n4 of “ha” and “ni”.

At this time, in the collating method according to the ninth method of the present invention, the character chain with the smaller number of two-character chains or the special two-character chain is searched with priority given in FIG. For example, (1) when the number of two-character chains "iro" is greater than the number of two-character chains "Iha", the two-character chain "Iha" is used as the first search character chain; The character chain "Iro" or (2) the two-character chain of the combination of the two-character chain preceding the special character, the first character of the two-character chain preceding the special character, and the character immediately following the special character It becomes a search character chain. Hereinafter, the collation of the character strings is performed in the same manner as in the fourth invention, except for the special two-character chain 2207 and the two-character chain 220.
Then, it is determined whether or not the number of appearances n3 of the second character "ha" in the special character chain 2207 matches the number of occurrences of the first character "ha" in the two-character chain 2208. Thus, the character strings are collated.

FIG. 21 shows the structure of a character string collating apparatus according to an embodiment of the ninth method of the present invention. In FIG. 21, reference numeral 2101 denotes a special character detector for detecting the special character "a" from the registered character string 2201;
Generates a two-character chain when there is no special character from the character string 2201, and detects a two-character chain 2202, 2204 registered in the two-character chain memory 2104 with the number of appearances of the character as a set. A special character chain 2203 that spans special characters is generated from the character string 2201.
A special two-character chain detector that registers the number of appearances of a character as a set in the two-character chain memory 2104, 2105 is a special character detector that detects the special character "a" from the character string 2205 to be searched, and 2106 is a special character detector that detects a special character from the character string 2201. If there is no character, a two-character chain is generated and the two-character chain 2202, 220
2 is a two-character chain detector for detecting the character string 2
The special two-character chain detector 2108 that generates a special character chain 2203 that straddles a special character from the special character chain 05 from the two-character chain detector 2106 and the special two-character chain detector 2107. The character string is detected from the character chain memory 2104, the continuity of the character chain is determined from the number of appearances of the character by the comparator 2108, and the matching of the search character string is determined by the control unit 2109.

Therefore, in this method, a specific special character "a" can be chained with the preceding and following characters irrespective of the number of appearances thereof, so that the character by the character chain is not restricted by the special character "a". Column matching can be performed. It is needless to say that in the case of a collation including a special character, for example, in the case of “a wa”, the collation of a character chain in which “ha” is the first character can be performed ignoring the special character.

(Embodiment 10) FIG. 23 shows a first embodiment of the present invention.
FIG. 24 is a block diagram of a character string collating apparatus according to the tenth embodiment, and FIG. 24 shows a concept of a character string collating method according to a tenth method of the present invention and a storage format of a recording medium storing full-text search data. I have.

In FIG. 24A, reference numeral 2401 denotes a character string “iro a ha ni” inputted at the time of registration, 2402 denotes a two-character chain “iro” registered first, and 2403 denotes a three-character chain next to 2402. 2404 is 2403 next to 2403
The character chain is "Hani". Here, "a" indicates a special character inserted into the character string to separate the meaning.

In FIG. 24 (c), reference numeral 2411 denotes a search character string “iro a ha ni” at the time of search, 2412 denotes a two-character chain “iro” to be searched first, and 2413 denotes a three-character chain “ro” next to 2412. a is ", and 2414 is a two-character sequence" Hani "following 2413.

In FIG. 24B, a two-character chain 2402
Is the appearance position n of "i", and the three-character chain 2403 is "ro"
The two-character chain 2404 stores the appearance position n + 2 of “ha”. Two-character chain 2402, 240
The 4 and 3 character chains 2403 are stored in different areas, and identify whether they are 2 character chains or 3 character chains. Search character string figure 24 (c)
In the collation method according to the tenth method of the present invention, a two-character chain 2402 corresponding to the "color" of the two-character chain 2412 is detected from the area storing the two-character chain, and the appearance position n And the three-character sequence 24 following 2412
The three-character chain 2403 corresponding to thirteen “ro aha” is detected from the area where the three-character chain is stored, and whether or not the appearance position n + 1 at this time coincides with the appearance position +1 of 2402 is determined. Judge. If they match, a two-character chain 2404 corresponding to the two-character chain 2414 “Hani” next to 2413 is detected from the area storing the two-character chain, and the appearance position n + 2 at this time is the appearance position of the aforementioned 2403 + 1
It is determined whether or not they match. If they match, the character string 211
Is determined to match 201. As described above, the character strings are collated.

FIG. 23 shows the configuration of a character string collating apparatus according to an embodiment of the tenth method of the present invention.

In FIG. 23, reference numeral 2301 denotes a two-character chain 2402, 240 to be registered from a character string 2401 to be registered.
4, and a two-character chain position detector 2302 for detecting their appearance position, a three-character chain 2403 to be registered from the registered character string 2401, and a detection for their appearance position 3
A character chain position detector 2303 is a two-character chain 2402,
Reference numeral 2404 denotes a two-character chain position memory for storing the appearance positions thereof, and reference numeral 2304 denotes a three-character chain position memory for storing the three-character chain 2403 and the appearance position of the chain.
Reference numeral 11 denotes a two-character chain detector that detects two-character chains 2412 and 2414 to be searched from a character string 2411 to be searched;
12 is a three-character chain 2 to be searched from the character string 211 to be searched
A three-character chain detector 231 for detecting 413 is a two-character chain 241 detected by the two-character chain detector 2311.
2, 2414 is detected by the two-character chain position memory 2303, or the three-character chain 2413 detected by the three-character chain detector 2312 is detected by the three-character chain position memory 2304, and the appearance position of each detected character chain is detected. The comparator 2314 determines whether or not matches the appearance position +1 of the character chain detected immediately before, and 2314 detects all of the 2 characters detected from the two-character chain detector 2311 and the three-character chain detector 2312.
The comparator 231 checks for a match on a character or a three-character chain.
The control unit determines in step 3 whether the character strings match.

In the character string collating device configured as described above, the two-character chain 2402 and 2404 in FIG. 24B is stored in the two-character chain position memory 2303, and the two-character chain 2403 in FIG. The following describes the operation performed when the three-character chain is stored, and “iroa wa ni” in FIG. 24C is input as a search character string.

When the search character string "color a han" is input, the two-character chain detector detects a two character chain "iro" and "hani" that do not include "a" specified as a special character in advance. Detected and output to the comparator 2313. The three-letter chain detector detects a three-letter chain “roaha” in which “a” is inserted, centering on “a” specified as a special character in advance, and outputs it to the comparator 2313.

At this time, the output to the comparator may be in the order of the sequence "color", "color a", "color", or together with the character chain information, "color", "color a", "color". May be output at the same time.

The comparator 2313 distinguishes between the output from the two-character chain detector and the output from the three-character chain detector.
"Roha" corresponding to "iro", "hani" and "roaha"
Are detected, and the chain is determined based on the number of appearances.

The comparator discriminates between a two-character chain and a three-character chain, and detects them from different chain memories, thereby performing a search by distinguishing between "iro a ha ni" and "iro ha ni" as character strings to be searched. Becomes possible.

As described above, according to the present embodiment, it is possible to avoid an increase in the chain memory due to special characters without being limited by the number of appearances of a specific special character “a” specified in advance. At the same time, it is possible to efficiently perform a chain extraction process based on the coincidence of the number of appearances.

In this embodiment, the special character is "a".
However, by replacing the special character sequence “a, a..., A” with “a”, a character that distinguishes whether a special character is inserted or not is not limited by the number of appearances of the special character. String matching by chaining can be performed.

That is, "Iro (one or more special characters) is a character"
And "irohani" can be searched using different search characters.

In this embodiment, different chain memories are provided to distinguish between a two-character chain and a three-character chain (insertion of special characters). By providing a displacement, for example, a two-character chain and a three-character chain can be stored as shown in FIG.
In this case, the appearance positions of the character chains 2402, 2403, 2404 are n, n + 1, n + 3, the displacements are 1, 2, 1, and
The continuity of each character chain is compared to determine whether the appearance position of each character chain matches the appearance position of the character chain immediately before the character chain plus the displacement, so that the two-character chain or the three-character chain is identified by the displacement. Then, by storing these data in the same area, the character string can be collated by the tenth method of the present invention.

(Embodiment 11) FIG. 26 is a conceptual diagram showing a configuration of a character string collating apparatus according to an eleventh embodiment of the present invention, and FIG. 25 is a character string collating method according to an eleventh method of the present invention. And the storage format of a recording medium storing full-text search data.

In FIG. 26A, reference numeral 2601 denotes a character string “iro a ha ni” inputted at the time of registration, and reference numeral 2602 denotes a specific special character “a” for the character string 2601 by a subsequent character “ha”. The character string "iro a1 ha ni" changed to "a1", 2603 is a two-character chain "iro" to be registered first, 2604 is a two-character chain "ro a1" next to 2603, and 2605 is 2604. The next two-letter chain "a1
"", And 2606 is a two-character chain "" in the next of 2605.

Here, "a" is a special character inserted into a character string to separate meanings and the like, and "a1" represents a specific symbol or code not to be searched.

In FIG. 26 (c), reference numeral 2611 denotes a search character string "iro a ha ni" at the time of search, and 2612 denotes a character string 261.
1 for the special character "a" followed by the character "wa"
The character string "iro a1" changed to "a1" uniquely determined by
"Hani", 2613 is a two-character chain "iro" to be searched first, and 2614 is a two-character chain "ro a" following the 2613.
1 ", 2615 is the two-letter chain following 2614" a1 ",
2616 is a two-character chain “Hani” next to 2615.

In FIG. 26B, a two-character chain 260
3 indicates the occurrence position n in the search target character string of "i"
The character chain 2604 stores the appearance position n + 1 of “ro”, the two-character chain 2605 stores the appearance position n + 2 of “a1”, and the two-character chain 2606 stores the appearance position n + 3 of “ha”.

At this time, in the collation method according to the eleventh method of the present invention, 2 characters corresponding to “color” in the two-character chain 2613 are used.
A character chain 2603 is detected, and a two-character chain 2601 corresponding to “a1” of a two-character chain 2614 next to 2613 is detected.
4 is detected, and it is determined whether or not the occurrence position n + 1 of the two-character chain 2604 at this time matches the occurrence position n of the detected two-character chain 2603 by +1. If they match,
Next, it is determined whether or not the value obtained by adding +1 to the appearance position n + 1 detected in 2604 matches the appearance position n + 2 of the two-character chain 2605 corresponding to “a1” in the two-character chain following 2614. If they match, then it is determined whether the occurrence position n + 2 detected at 2605 is equal to the +1 value, and whether the occurrence position n + 3 of the two-character chain 2606 corresponding to “Hana” of the next two-character chain after 2615 matches. If they match, the character string 2611
Is determined to match 2601. As described above, the character strings are collated.

FIG. 25 shows the configuration of a character string collation apparatus according to an embodiment of the eleventh method of the present invention.

In FIG. 25, reference numeral 2501 denotes a character string 2401 in which a character string 2601 to be registered is changed from a specific special character “a” to “a1” uniquely determined by the subsequent character “ha”.
A character string converter for changing to 602, and 2502 for the character string 26
2 character chain 2603, 2604, 26 to be registered from 02
05, 2606 and a two-character chain position detector for detecting the appearance position of the two-character chain, and a two-character chain position memory 2503 for storing the two-character chain 2603, 2604, 2605, 2606 and the occurrence position of the character chain , 2504 designates a character string 2611 to be searched as “a” in which a specific special character “a” is uniquely determined by a subsequent character “ha”.
A character string converter for changing to a character string 2612 changed to "1", 2505 is a two-character chain detector for detecting a two-character chain 2613, 2614, 2615, 2616 to be searched in the character string 2612, and 2506 is a two-character chain detector 25
05 two-character chain 2613, 2614, 2
The comparator 2507 detects 615 and 2616 in the two-character chain position memory 2503, and determines whether or not the detected occurrence position of the two-character chain matches the appearance position of the two-character chain detected immediately before by +1. Two-character chain detector 2505
The comparator 250 for all two-letter chains detected from
The control unit determines in step 6 and determines whether the character strings match.

The operation of the thus constructed character string collating apparatus will be described. When the registered character string is input, the character string conversion unit 2501 converts the special character “a” specified in advance into symbols or codes that are not to be searched by predetermined characters, that is, symbols other than the search character string, Convert to code and output.

As shown in FIG. 26 (d), the character string conversion means stores the correspondence corresponding to the character following the special symbol and the conversion to which symbol. This correspondence is 2621, 26
Even if the correspondence differs for each character, such as 22,
May correspond to a group of characters.

In the converted character string, a two-character chain and its appearance position are detected by a two-character chain detector as in the tenth embodiment, and are stored in a two-character chain position memory.

On the other hand, when the search character string is given, the character string converter 2504 converts the special character into a symbol or code other than the search character string according to the same correspondence as that used in the character string conversion 2501, and converts the special character into two characters. Output to the chain detector. 2
The character chain detector detects the two-character sequence, and outputs the result to the comparator 2506.
Output to

The comparator 2506 detects a match between character chains in accordance with the contents of the two-character chain memory according to the same procedure as in the tenth embodiment. However, in the eleventh embodiment, it is not necessary for the comparator to distinguish between a two-character chain and a three-character chain as in the tenth embodiment.

As described above, according to the present embodiment, it is possible to avoid an increase in the chain memory of the special character “a” having a high appearance frequency, and to replace the same special character with a plurality of different characters in accordance with the subsequent characters. By converting to characters, candidates for checking the coincidence of the number of appearances for extracting a chain are dispersed into a plurality of candidates, so that the processing time is reduced.

In the present embodiment, the special character "a"
Is determined based on the character after that, but it is clear that the same effect can be obtained even when the conversion destination is determined based on the character before the special character.

The schematic diagram of the case where the present invention is implemented as a computer is the same as that of FIG. 1A. In this embodiment, the two-character chain position memory 2503 corresponds to the external recording device 40.

In this embodiment, a two-character chain position memory as shown in FIG. 26B is provided.
By providing a displacement as character chain position information, for example, FIG.
As shown in (e), a two-character chain can be stored. In this case, the character chains 2603, 2604, 2605, 26
Assuming that the appearance position of 06 is n, n + 1, n + 2, n + 3 and the displacement is 1, 1, 1, 1, the continuity of each character chain is represented by the appearance position of the character chain immediately before the character chain + By comparing with or comparing with the displacement, the eleventh aspect of the present invention
By the method described above, character string collation can be performed.

(Embodiment 12) FIG. 27 is a block diagram showing a configuration of a character string collating apparatus according to a twelfth embodiment of the present invention, and FIG. 28 is a concept of a twelfth method of character string collation of the present invention. , And the storage format of the recording medium storing the full-text search data.

In FIG. 28 (a), reference numeral 2801 denotes a character string “iro a ha ni” inputted at the time of registration, and 2802 denotes a specific special character “a” with respect to the character string 601; "Roro", which consists of "ro" uniquely determined by "ro" and "ro", and the subsequent character "ha", which consists of "ha" and "ha", which is uniquely determined by "ha" Character string changed to 'ha'
Is' Hani ', 2803 is the two-letter chain "iro" that is registered first, 2804 is the two-letter chain "Roro" next to 2803, and 2805 is the two-letter chain "Ro'" next to 2804.
"", 2806 is the two-letter chain following 2805 ""
"Ha", 2807 is the next two-character chain "hani" after 2806.

Here, “a” is a special character inserted into a character string for separating meanings, etc.
Represents a specific symbol or code not to be searched.

In FIG. 28 (c), reference numeral 2811 denotes a search character string "iro a ha ni" at the time of search, and 2812 denotes a character string 281.
The special character "a" is replaced by "ro" consisting of "ro" which is uniquely determined by "ro" and "ro", and the subsequent character "ro". "Ha" is a character string changed to "ha'ha" consisting of "ha '" and "ha" uniquely determined by "ha".
Is 'Hani', 2813 is the two-letter chain “iro” that is searched first, 2814 is the two-letter chain “Roro” next to 2813, and 2815 is the two-letter chain “Ro” next to 2814.
"', 2816 is the two-letter chain following 2815""
"Ha", 2817 is the two-character chain "hani" following 2816.

In FIG. 28B, a two-character chain 280
3 is an appearance position n of "iro", a two-character chain 2804 is an appearance position n + 1 of "roro '", a two-character chain 2805 is an appearance position n + 2 of "ro'ha'", and a two-character chain 2806.
Is the appearance position n + 3 of "ha'ha" and the two-character chain 2807
Stores the appearance position n + 4 of "Hani".

At this time, in the collation method according to the twelfth method of the present invention, two characters corresponding to “color” of the two-character chain 2813
The character chain 2803 is detected, and + appears at the appearance position n of 2803.
The two-character chain 2804 corresponding to “Roro '” in the two-character chain 2814 following the 2813 is detected, and the value 280 is detected.
It is determined whether or not the appearance position n + 1 of 4 matches. If they match, then the value obtained by adding +1 to the appearance position n + 1 detected in 2804 and the two-character chain 2805 corresponding to “ro'wa '” in the next two-character chain following 2814 are detected, and the appearance position n + 2 of 2805 matches. It is determined whether or not to do. If they match, then a value obtained by adding +1 to the appearance position n + 2 detected in 2805 and 2
A two-character chain 2806 corresponding to “ha'ha” of the two-character chain following 815 is detected, and it is determined whether or not the appearance position n + 3 of 2806 matches. If they match, then a value obtained by adding +1 to the appearance position n + 3 detected in 2806 and a two-character chain 2807 corresponding to “Hana” in the next two-character chain after 2816 are detected, and is the appearance position n + 4 of 2807 the same? Determine whether or not. If they match, it is determined that the character string 2811 matches 2801. As described above, the character strings are collated.

FIG. 27 shows the configuration of a character string collating apparatus according to an embodiment of the twelfth method of the present invention.

In FIG. 27, reference numeral 2701 denotes a specific special character “a” for a character string 2801 to be registered, and the preceding character “ro” is composed of “ro” and “ro ′” uniquely determined by “ro”. A character string converter that changes the character string 2802 to “roro '” and the subsequent character “ha” uniquely determined by “ha”. , 2702 are two-character chains 2803, 2804, 2805, 2 registered from the character string 2802.
2 to detect 806, 2807 and their appearance position
A character chain position detector, 2703 is a two-character chain 2803,
A two-character chain position memory for storing 2804, 2805, 2806, 2807 and their appearance positions, 2704
Replaces the character string 2811 to be searched for a specific special character "a" with the preceding character "ro" as "roro" consisting of "ro" uniquely determined by "ro" and "ro", and thereafter The character "wa" is a character string converter that is uniquely determined by "ha" and is changed to a character string 2812 which is changed to "wa'ha" consisting of "ha" and "ha". Character chains 2813, 2814, 281
A two-character chain detector that detects 5, 2816, 2817;
2706 is the value of 2 detected by the two-character chain detector 2705.
Character chains 2813, 2814, 2815, 2816, 2
817 is detected by the two-character chain position memory 2703, and a comparator for determining whether or not the detected occurrence position of the two-character chain matches a value obtained by adding +1 to the occurrence position of the two-character chain detected immediately before, and 2707 is a two-character comparator. The comparator 2706 determines all two-character chains detected from the chain detector 2705,
This is a control unit that determines whether the character strings match.

As means for converting a special character into a character uniquely determined by adjacent characters before and after the special character in the present invention, as shown in FIG. It is converted or its correspondence is stored. This correspondence may be different for each character, such as 2821 and 2822, or may correspond to a group of characters, such as 2823.

As described above, according to the present embodiment, it is possible to perform character string collation by character chain without being limited by the number of appearances of the special character “a”.

That is, according to the embodiment, the special character "a"
Is converted into separate characters by the characters before and after it, and the number of appearances of the converted characters is recorded.
By distributing the two-character chain file more finely, it is possible to avoid an increase in the chain memory for the special character “a” with a high appearance frequency of the frequently used special character. Efficiency can be improved.

The schematic diagram of the case where the present invention is implemented as a computer is the same as that of FIG. 1A. In this case, the two-character chain memory 2703 corresponds to the external recording device 40.

In this embodiment, a two-character chain position memory as shown in FIG. 28B is provided.
By providing a displacement as character chain position information, for example, FIG.
As shown in (e), a two-character chain can be stored. In this case, the character chains 2803, 2804, 2805, 28
06, 2807 are n, n + 1, n + 2, n +
Assuming that 3, n + 4 and the displacement are 1, 1, 1, 1, and 1, the continuity of each character chain is compared to see if the appearance position of each character chain matches the appearance position + displacement of the character chain immediately before the character chain. Thus, the character string can be collated by the twelfth method of the present invention.

(Embodiment 13) FIG. 29 is a block diagram showing a configuration of a character string collating apparatus according to a thirteenth embodiment of the present invention, and FIG. 30 is a concept of a thirteenth method of character string collating according to the present invention. , And the storage format of the recording medium storing the full-text search data.

FIG. 30 shows the concept of a character string collation method according to the thirteenth method of the present invention. In FIG. 30 (a),
3001 is a character string "Iro a Hani" input at the time of registration,
3002 is a two-character chain “iro” that is registered first, 3
003 is the second character of the three-character chain "roaha" in which the special character next to 3002 is inserted, and the three-character chain "rohaha" obtained by converting the special character "a" to the next third character "ha". , 3004 is a two-character chain "roha" of the first character and the second character of the three-character chain 3003, and 3005 is a first character of the second character and the third character of the three-character chain 3003 corresponding to the special character "a". Is a special two-character chain “Hana”, and 3006 is a two-character chain “Hani” next to 3005. In FIG. 30C, 3
011 is a search character string at the time of search “Iro a Hani”, 3012
Is the first two-character chain "iro" to be searched, and 3013 is 3
012 with the special character next to it inserted
The special character “a” is the second character of “ha” and the next third character is “ha”
The three-letter chain “Rohah” converted to “3,” 3014 is the two-letter chain “Rohah” with the first and second characters of the three-letter chain 3013, and 3015 is the second and third characters of the three-letter chain 3013.
Special 2 in which the first character is the special character "a"
The character chain “Hana” and 3016 are the two-character chain “Hani” next to 3015.

In FIG. 30B, a two-character chain 300
2 is the appearance position n of “iro”, the two-character chain 3004 is the appearance position n + 1 of “iroha”, and the special two-character chain 3005
Stores the appearance position n + 2 of “haha” in another area, the next two-character chain 3006 of the special two-character chain stores the appearance position of “hana”, and the same value n + 2 as the appearance position of the special two-character chain 3005. .

At this time, in the collation method according to the thirteenth method of the present invention, 2 characters corresponding to “color” in the two-character chain 3012 are used.
A character chain 3002 is detected, and +
It is determined whether or not the value obtained by 1 and the appearance position n + 1 of the two-character chain 3004 corresponding to “roha” of the first two-character chain 3014 of the three-character chain 3013 following the 3012 match.
If they match, then the value obtained by adding +1 to the appearance position n + 1 detected in 3004 and the special two-character sequence 3015 next to 3014
It detects that the appearance position n + 2 of the special two-character chain 3005 corresponding to “haha” matches. Next, it is determined whether or not the appearance position n + 2 of 3005 and the appearance position of the two-character chain 3006 corresponding to “Hana” in the two-character chain 3016 next to 3015 match. If they match, it is determined that the character string 3011 matches 3001. As described above, the character strings are collated.

FIG. 29 shows the structure of a character string collating apparatus according to an embodiment of the thirteenth method of the present invention.

In FIG. 29, reference numeral 2901 denotes a three-character chain of special characters for the character string 3001 to be registered.
003 and two-character chains 3002, 3006,
In the case of the three-character chain 3003, the special character of the second character is changed to the third character.
A three-character chain detector that converts the character to the same character and makes the character position of the second character the same as the character position of the third character;
902 is a two-character chain 3002 input from 2901;
A two-character chain position detector that detects the appearance position of 3006;
2903 is a three-character chain 3003 input from 2901
Two-character chain 3004 consisting of the first and second characters and special two-character chain 3005 consisting of the second and third characters
Special two-character chain generator 2904 for detecting two character sequences and the appearance position of each character chain.
2, 3004, 3006 and a two-character chain position memory for storing the appearance positions of their character chains, 2905 is a special two-character chain 3005 and a special two-character chain position memory for storing the appearance positions of the character chains, and 2911 is a search. For the character string 3011, the second character identifies a special character three-character chain 3013 and two-character chains 3012 and 3016. In the case of the three-character chain 3013, the special character of the second character is converted to the same character as the third character. Three-letter chain detector, 29
12 is a two-character chain 3012 input from 2911, 3
Two-character chain detector that detects 016, 291 is 291
The first and second characters of the three-character chain 3013 input from
A special two-character chain generator 2914 for detecting two two-character chains, that is, a two-character chain 3014 consisting of characters and a special two-character chain 3015 consisting of a second character and a third character, and 2914 is detected by the two-character chain detector 2912. Two character chain 301
2, 3014, and 3016 are detected by the two-character chain memory 2904, or the special two-character chain 3015 generated by the special two-character chain generator 2913 is detected by the two-character chain memory 2904. In the case of a special two-character chain, its appearance position matches the appearance position of the character chain detected immediately before, or when it is not a special two-character chain, its appearance position is the value obtained by adding +1 to the appearance position of the previously detected character chain. A comparator for determining whether or not
Reference numeral 2915 denotes a control unit which determines whether the two-character chains detected by the two-character chain detector 2912 and the special two-character chain detector 2913 coincide with each other by the comparator 2914, and judges character string coincidence.

Therefore, at this time, it is possible to perform character string collation by character chain without being restricted by the appearance of the specific special character “a”.

In this embodiment, a different chain memory is provided to distinguish between a two-character chain and a special two-character chain (special character insertion) as shown in FIG. 30B. By providing a displacement for identifying a chain or a special character chain, a two-character chain and a special two-character chain can be stored, for example, as shown in FIG. In this case, character chain 3
The appearance positions of 002, 3004, 3005, and 3006 are n, n + 1, n + 2, n + 2, and the displacement is 1, 1, 0, 1, and the continuity of each character chain is determined by the appearance position of each character chain immediately before the character chain. By comparing with the appearance position of the character chain + displacement, whether the two-character chain or the special two-character chain is identified by the displacement, these data are stored in the same area, and the fifteenth aspect of the present invention is described. By the method described above, character string collation can be performed.

(Embodiment 14) FIG. 32 shows the concept of a fourteenth character string registration method for character string collation according to the present invention.

In FIG. 32 (a), reference numeral 3201 denotes a registered character string "color a wa ni b color a ha ni" inputted at the time of registration.
“A” and “b” are special characters, and the document number is N. Numbers the first registered character string. 3
Reference numeral 216 denotes a character position and a unique number of the registered character string 3201, and the special character “a” is set with n as the head of the registered character string.
Numbering is performed in ascending order excluding "b", the character position is n for the first character "i", the fourth character "ha" is n + 2, and so on, except for special characters. . Also, unique numbers are assigned to the special characters “a” and “b”,
“A” is numbered m, and “b” is numbered l. Next, a two-character chain is created. A character chain that does not include the special characters “a” and “b” in the registered character string 3201 is searched.
Character chain "Iro" 3202, two-character chain "Hani" 320
3 and 2 character chain "Iro" 3204 and 2 character chain "Hani"
3205 is created. Subsequently, a special two-character chain “roa” 3206 including special characters and a special two-character chain “aha” 32
07 and the special character "ro" 3208 combining the character "ro" immediately before the special character "a" and the character "ha" immediately after the special character "a".
Create Similarly, a special two-character chain “nib” 32 is applied to the sixth special character “b” of the registered character string.
09, “b” 3210, “ni” 3210, a special two-character chain “ro a” 3212, “a wa” 3213, “roha” for the ninth special character “a” in the registered character string. "32
13 is created.

FIGS. 32 (b) and (c) show character chain information created from the two-character chain and the special two-character chain created in FIG. 32 (a). The character chain information includes a two-character chain or a special two-character chain, a character position, and a document number. The character chain information is arranged for each character type of the first character in the character chain and in document number order. First, FIG. 32 (b) will be described. For the two-character chain “Iro”, 3202 and 3204 are arranged from FIG. 32 (a), and the character positions of each two-character chain are n and n + 1. Similarly, for the two-character chain “Hani”, 3203 and 3205 are arranged and the character positions are n + 2 and n +
It becomes 6. For a special two-character chain that includes or sandwiches the special character “a”, the second character “R” in the registered character string
Special two-letter chain "roha" as the first character in the character chain
3208, and special two-character chain 320 immediately after "ro"
9 are taken out and arranged continuously. At this time, the character position is "roha" with n + 1, and "roa" with m. Similarly, the special character chain 3 is applied to the eighth character “R” of the registered character string.
It is set in the order of 214 and 3212. If the first character of the special two-character chain is "a", the character position of the second character is assigned.
Assigned in the order of the character positions of 07 and 3213. Subsequently, the character chain information for the special character "b" is created in the order of the special two-character chain "ni" 3211 and "ni b" 3209 with "ni" as the first character, as in FIG. 32 (b). Set, and use the special two-character chain "b" with "b" as the first character.
“3210” is set.

At this time, the collating method according to the fourteenth method of the present invention will be described with reference to FIG. With respect to the search character string 3217 "iro a ha ni", first, the presence or absence of the special character "a" is checked from the top, and if it is not included, a two-character chain "iro" 3218 is created. Continue with "ro a" and "a
Is created and detected as special two-character chains 3220 and 3221. At this time, the special character "a" is detected between the special character "a" and the special character "a". Subsequently, a two-character chain “Hani” 3222 is detected.

Next, a character chain corresponding to the detected two-character chain and the special two-character chain is extracted from the character chain information in FIG. 32 (b). The character chain corresponding to the two-character chain “Iro” 3218 includes 3202 and 3204 in FIG.
First, 3202 is taken out. Subsequently, the character chain corresponding to the special two-character chain "roha" 3219 is 32 in FIG.
08, 3214, and first retrieves 3208. 3
The document numbers of 202 and 3208 are both N and match,
Since the character positions of the two character chains are consecutive character positions of n and n + 1, it is determined that 3202 and 3208 are continuous. Subsequently, the character chain following the special two-character chain 3208 is examined in the character chain information of FIG. 32 (b), and the character chain 3206 “a” is extracted. 32
06 is the first character "ro" and the second character is the special character "a"
Since the value of the character position is the unique value m and the document number M, it is determined to be a special two-character chain following the special two-character chain "roha" 3208.

Next, the character chain information corresponding to the special two-character chain “a wa” 3221 is checked from FIG.
The character chain “a wa” 3207 is extracted. Since the document number of 3207 is N and the character position is n + 2, it is determined to be a special two-character chain following the character position n + 1 of the special two-character chain "roha" 3208.

Next, character chain information corresponding to the two-character chain “Hana” 3222 is examined from FIG. 32 (b), and the two-character chain “Hana” 3203 is extracted. Since the document number of 3203 is N and the character position is n + 2, it is determined that it is a two-character chain following the character position n + 1 of the special two-character chain "roha" 3208. As described above, it is determined that the search character string 3217 is included in the registered character string 3201.

In the above collation method, when the character chain information shown in FIG. 32B corresponding to the two character chain 3218 is extracted, 3204 out of the two character chains 3202 and 3204 is the same as the collation method. By the way,
Two-character chain "iro" 3204 (character position n + 4, document number N), special two-character chain "iro" 3214 (character position n
+5, document number N), special two-character chain "ro a" (character position m, document number N), special two-character chain "a" (character position n + 6, document number N), two-character chain "hani" (Character position n + 6, document number N) is detected, and it can be determined from the comparison of continuity between the document number and the character position that they match. From the above, the search character string 3217 is the registered character string 3
201 is determined to be included.

A search character string 3223 containing a special character "b" in the search character string "Hanib Iro" is then obtained by the same collation method.
Then, character chain information is obtained from FIGS. 32 (b) and 32 (c), and matching of document numbers and collation of continuity of character positions are performed. For the two-character chain "Hani" 3224, the two-character chain "Hani" 3203 (character position n + 2, document number N) and for the special two-character chain "Nii" 3225, the special 2
Character chain "ni" 3211 (character position n + 3, document number N), special 2 character chain "ni b" 3226
Character chain "ni b" (character position 1, document number N), special 2
A special two-character chain "b"
(Character position n + 4, document number N) and two-character chain "iro" (character position n)
(+4, document number N), and determines that the search character string 3223 is included in the registered character string 3201.

FIG. 31 shows the structure of a character string collating apparatus according to an embodiment of the fourteenth method of the present invention.

In FIG. 31, reference numeral 3101 denotes a special character detector which detects a specific special character “a” in the registered character string 3201 and gives the character position 3216 of the registered character string.
02 is a two-character chain 3202, 3 from the registered character string 3201.
203, 3204, and 3205, a two-character chain encoder for creating a character position and a document number of a two-character chain, and 3103 a special two-character chain 3206 and 320 from a registered character string 3201.
7, 3208, 3209, 3210, 3211, 321
2, 3213, 3214, a special two-character chain encoder for creating the character position and the document number of the special two-character chain, and 3104, a two-character chain encoder 3102 and the special two-character chain encoder 32
This is a character chain combination judging unit that generates the character chain information shown in FIGS. 32 (b) and (c) from the two-character chain, special two-character chain, character position, and document number created in step 03 and stores it in the two-character chain memory 3105. .

Reference numeral 3106 denotes search character strings 3217 and 3223
A special character detector 3107 for detecting a special character "a" or "b" from a two-character chain 3218, 3222, 32
Two-character chain detector for detecting 24, 3228, 3208
Is a special two-character chain 3219-3221, 3225-32
The special two-character chain detector 3109 for detecting the character string 27 determines the collation order from the two-character chain and the special two-character chain. In the case of the search character string 3217, 3218, 3219, and 322
0, 3221 and 3222, in order of the search character string 3223
In the case of, 3224, 3225, 3226, 3227,
A character chain combination determiner 3228 also arranges the character chains in order. The character chain 3210 fetches two character chains corresponding to the character chain sent from the 3109 character chain combination determiner from the two-character chain memory 3105 in order, two by two. A comparator that extracts the document number and character position of one character chain and sends the data to the 3111 control unit, 3111 checks the continuity of the character chain from the data sent from the comparator 3110,
If they are consecutive, the data of the next character chain is compared with the data in the comparator 3110.
And ends the collation if they are not continuous.

Therefore, in this method, a specific special character “a” can be chained with the preceding and following characters irrespective of its appearance frequency, so that the number of appearances of the special character “a” is not restricted. String matching by chaining can be performed. In addition, collation that includes special characters at the beginning,
For example, in the case of "a", ignore special characters and
It is needless to say that the collation can be performed at high speed by performing the collation of the character chain in which is the first character.

(Embodiment 15) FIG. 34 shows the concept of a fifteenth character string registration method for character string collation according to the present invention.

In FIG. 34 (a), reference numeral 3401 denotes a registered character string "iro a wa ni ni" inputted at the time of registration.
In the registered character string 3401, “a” is a special character, and the document number of the registered character string is M. 3402 is the registered character string 34
The special character "a" is removed from the character 01, and the character "ha" immediately after the special character is replaced with the character "ha *" immediately after the special character.
Is a registered character string marked as. Reference numeral 3409 denotes registered character positions in which a character position at the head of the registered character string 3401 is set to n and numbered sequentially except for the special character “a”. First, in step 3402, a two-character chain is created except for the character "ha *" immediately after the special character. 3403 is a two-character chain of “iro” and a two-character chain of “ni” 340
5. A two-character chain 3406 of “Hani” is created. Next, the two-character chain “Iha *” 3 of two characters before the special character “a”, that is, two characters before “ha *” and “ii” and “ha *”.
407, a two-character chain of “ro” and “ha *”, which is the character preceding “ha *”, and a special two-character chain of “roha *” 3408, and “ha *” and the next character “ni” "Ha *" is created.

FIG. 34 (b) shows the configuration of character chain information created from the two-character chain and the special two-character chain created in FIG. 34 (a). The character chain information includes a document number, a two-character chain, a special two-character chain, a character position, and a special two-character chain flag. The two-character chain and the special character chain are consecutively arranged for each character type of the first character. Special 2
The starting position of the character chain is indicated by a special two-character chain flag. In FIG. 34 (b), the character chain information 3411 is
For the first character of the character chain "ha" and "ha *" in which the special character "a" is inserted immediately after "ha", the document number 341
2, two-character chain 3413 including "ha", character chain information at character position 3414 of "ha", and document number 3412,
A special two-character chain 3415 including “ha *”, character chain information indicating the character position (n) 3416 of “ha *”, and a special two-character chain flag 34 indicating the position of the special two-character chain 3415
17. Here, there are N two-character chains including "wa" as the first character, and the special two-character chain follows immediately after the two-character chain.
17 stores N + 1 which is the start position of “*”. If the second character is the character immediately after the special character “a”, it is assumed that the character is included in the two-character chain “was X” (X is the second character).

FIG. 34 (c) shows an example of storing character chain information in the case of FIG. 34 (a). 3418 is 2
The first character of the character chain is the character chain information 3419 of “I”, which is stored in the two-character chain 3419 of “Iro” and the two-character chain “Iha *” 3420 with “I” as the first character. . In this case, the value of the special two-character chain flag is “0” because the first character does not include the character immediately after the special character “a”. Similarly, a two-character chain is stored in the character chain information 3422 having "ro" as the first character, a value "0" is stored in the special two-character chain flag 3423, and "ni" is the first character. Is stored in the character chain information 3428, and the value “0” is stored in the special two-character chain flag 3429. On the other hand, the character chain information 3424 in which the first character is “Hana” is a two-character chain “Hani” 34 25
And “n + 4”, which is the character position of “ha”, are stored first, followed by a special two-character chain 3426 with “ha *” as the first character and “n + 2”, which is the character position of “ha *”. Stored as character chain information. Further, since the special two-character chain flag of “ha *” is the second character chain information having “ha” as the first character, the value “2” is stored.

The collation method according to the fifteenth method of the present invention will be described with reference to FIG. No. 35 (a)
Indicates a collation method in the case of a search character string including the special character "a" between them, and the second character from the beginning includes the special character. With respect to the search character string 3501 “ro a ha ni”, a character “ha” following the special character “a” is set to “ha *” to create a search character string 3502, and a two-character chain 3503
"Roha *" and special character chain 3504 "Ha * ni" are detected. Subsequently, a two-character chain whose first character is "ro" is searched from the character chain information in FIG. 34 (c), and a two-character chain "roha *" is detected from the character chain information 3422. At this time, the character position n + 1 of "roha *" and the document number M are extracted and stored. Next, the special two-character chain 3504 “Ha * ni” is obtained from the character chain information in FIG. 34 (c. The first character of the special two-character chain “Ha * ni” is “Ha *”, so the first character Detected from the character chain information 3424 of “ha”,
Since the position of "ha *" is "2" from the special two-character chain flag 3427, the special two-character chain 3426 is detected. At this time, the document number of 3426 is M, and the character position is n +
Since it is 2, the two-character chain "roha *" detected earlier
Character number n + 2 and character position n + 2 next to character position n + 1.
”Is determined to be continuous, and the search string“ roa han ”
Is determined to be included in the registered character string. Thus, the character strings are collated.

Next, FIG. 35 (b) shows a collation method in the case of a search character string including the special character "a" at the head.
For the search character string 3505 “a wa ni”, first, the character “ha” following the special character “a” is set to “ha *” and the search character string 3
506 is created, and the special character chain 3507 "Ha * ni" is detected. Subsequently, a special two-character chain in which the first character is "ha *" is searched from the character chain information in FIG. 34 (c), and the character string collation of the special two-character chain "ha * ni" in FIG. 35 (a) is searched. It is determined whether or not the character chain information is continuous by the same method as that described above.
426 is detected. Thus, the character strings are collated. If the search character string is “a wa” 3508, a search character string 3509 is created and a special two-character
0 is detected. In this case, since the second character of the special two-character chain does not exist, when detecting the character chain from the character chain information, if there is character chain information in which the first character is "ha *", it is detected. Judge that it was done.

Next, FIG. 35 (c) shows a method of collating a character string in the case of a search character string including the special character “a” between the first and third characters from the beginning. I have.
In response to the search character string 3511 "iroa ha ni", the character immediately after the special character "a" is replaced with "ha *" to create a search character string 3512. Next, the two-letter chain “iro” 3513 that does not include the special character “a”, the two-letter chain “Iha *” 3514, and the special two-letter chain 3515 “Ha * ni” are detected. Subsequently, a two-character chain “IRO” whose first character is “I” is searched from the character chain information in FIG.
From the character chain information 3418, a two-character chain "iro" 3419
Is detected. At this time, the character position n of "iro" and the document number M are extracted and stored. Next, a two-character chain 351
4 "I *" is obtained from the character chain information in FIG. 34 (c). Since the first character of the two-character chain “I *” is “I”, it is detected from the character chain information 3418, and the two-character chain 3420 of the second character “HA *” is detected. At this time, 342
Since the document number of 0 matches M and the character position matches n,
It is determined that the previously detected two-character chain “Iro” is continuous, and that “Iro a” is included in the registered character string of the document number M. Next, the two-character chain "I *" 3514
And the special two-character chain “ha * ni” 3515 are compared. This collation is the same processing as in the 35th (a), but note that the difference between the character positions of “i” and “ha *” is 2, and the two-character chain 3420 of “ii *” Since the character position of the special two-character chain 3426 of the character position n and “ha * ni” is n + 2 and the difference between the character positions is 2, the character chains “ii *” and “ha * ni” are consecutive. It is determined that the search character string is included in the registered character string. Thus, the character strings are collated.

FIG. 33 shows the structure of a character string collating apparatus according to an embodiment of the fifteenth method of the present invention.

In FIG. 33, reference numeral 3301 denotes a special character detector that detects a specific special character “a” in a character string 3401 to be registered and gives a character position 3410 of a registered character position 3409; A two-character chain 3403, 3405, 340 from a registered character string 3402 in which the character "ha" immediately after the special character "a" is a special character.
6, 3407, 3408 and a special character chain 3404 determine a character chain combination determining unit 3303, which creates a pair of a two-character chain and a document number, and converts the character chain information into two.
A two-character chain encoder stored in the character chain memory 3305;
Reference numeral 3304 denotes a special two-character chain encoder for creating a set of a special character chain and a document number, and storing character chain information and a special character chain flag in a two-character chain memory 3305. Character strings 3501, 3505, and 3508 to be searched for. , 351
A special character "a" is detected from the search character string 350
2, 3506, 3509, 3512 are special character detectors for generating 3507, and 2307 is a two-character chain 3503, 3507, 3
510, 3513, 3514 and special two-character chain 35
A character chain combination determiner 3308 for determining the combination of 04, 3515 is a two-character chain 3503, 3507, 351
A two-character chain detector that creates 0, 3513, 3514,
Reference numeral 3509 denotes a special two-character chain detector for generating special two-character chains 3504 and 3515, and 3510 denotes 3508 and 09.
351 retrieves a two-character chain and a special two-character chain that match the two-character chain and the special two-character chain from the two-character chain memory 3505, and compares the character continuity.
Reference numeral 1 denotes a control unit that stops collation if the character chains collated by the comparator 3510 do not match, and instructs the comparator 3510 to collate with the next character chain if they match, and determines a collation of character continuity. is there.

Therefore, in this method, since the specific special character "a" can be linked with the preceding and following characters regardless of the frequency of occurrence, the character can be generated without being limited by the number of appearances of the special character "a". String matching by chaining can be performed. In addition, collation that includes special characters at the beginning,
For example, in the case of "a", ignore special characters and
It is needless to say that the collation can be performed at high speed by performing the collation of the character chain in which is the first character.

[0203]

As described above, according to the present invention, when a character string to be collated includes a special character having a high appearance frequency, the character chain including the special character ignores the appearance frequency of the special character. Therefore, it is possible to easily cope with character string collation of a language including such characters, and the effect is large.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram of a character string collation device according to a first embodiment of the present invention.

FIG. 2 is a conceptual diagram showing a character string collation method according to the first embodiment of the present invention.

FIG. 3 is a block diagram of a character string collating apparatus according to a second embodiment of the present invention.

FIG. 4 is a conceptual diagram showing a character string collation method according to a second embodiment of the present invention.

FIG. 5 is a block diagram of a character string collating apparatus according to a third embodiment of the present invention.

FIG. 6 is a conceptual diagram showing a character string collation method at the time of registration according to a third embodiment of the present invention.

FIG. 7 is a conceptual diagram illustrating a character string matching method based on the number of appearances of a two-character chain according to a third embodiment of the present invention.

FIG. 8 is a conceptual diagram showing a method of collating a character string at the time of retrieval according to a third embodiment of the present invention.

FIG. 9 is a block diagram of a character string collating apparatus according to a fourth embodiment of the present invention;

FIG. 10 is a conceptual diagram showing a fourth character string collation method of the present invention.

FIG. 11 is a block diagram of a character string collating apparatus according to a fifth embodiment of the present invention.

FIG. 12 is a conceptual diagram showing a fifth character string collation method of the present invention.

FIG. 13 is a block diagram of a character string collating apparatus according to a sixth embodiment of the present invention.

FIG. 14 is a conceptual diagram showing a sixth character string collation method of the present invention.

FIG. 15 is a conceptual diagram showing a seventh character string collation method of the present invention.

FIG. 16 is a block diagram of a character string collating apparatus according to a seventh embodiment of the present invention;

FIG. 17 is a conceptual diagram showing an eighth two-character chain data creation method and a character string collation method according to the present invention.

FIG. 18 is a flowchart of creating two-character chain data in an embodiment of the eighth method of the present invention.

FIG. 19 is a flowchart of search string matching in an embodiment of the eighth method of the present invention.

FIG. 20 is a block diagram of a character string collating apparatus according to an eighth embodiment of the present invention.

FIG. 21 is a conceptual diagram showing a ninth two-character chain data creation method and a character string collation method according to the present invention.

FIG. 22 is a block diagram of a character string collating apparatus in a ninth embodiment of the present invention.

FIG. 23 is a conceptual diagram of a character string collation device according to a tenth embodiment of the present invention.

FIG. 24 is a conceptual diagram showing a character string collation method according to a tenth method of the present invention.

FIG. 25 is a block diagram of a character string collating apparatus according to an eleventh embodiment of the present invention.

FIG. 26 is a conceptual diagram showing a character string collation method according to an eleventh method of the present invention.

FIG. 27 is a block diagram of a character string collating apparatus according to a twelfth embodiment of the present invention.

FIG. 28 is a conceptual diagram showing a character string collation method according to a twelfth method of the present invention.

FIG. 29 is a block diagram of a character string collation apparatus according to a thirteenth embodiment of the present invention.

FIG. 30 is a conceptual diagram showing a character string collation method according to a thirteenth method of the present invention.

FIG. 31 is a block diagram of a character string collating apparatus in a fourteenth embodiment of the present invention.

FIG. 32 is a conceptual diagram showing a character string registration method of the character string collation device in the fourteenth embodiment of the present invention.

FIG. 33 is a block diagram showing a character string collating apparatus according to a fifteenth embodiment of the present invention;

FIG. 34 is a conceptual diagram showing a method for registering a character string in the character string collating apparatus according to the fifteenth embodiment of the present invention.

FIG. 35 is a conceptual diagram showing a character string collation method of the character string collation device in the fifteenth embodiment of the present invention.

FIG. 36 is a block diagram of a conventional character string collation device.

FIG. 37 is a conceptual diagram showing a conventional character string collation method.

[Explanation of symbols]

 Reference Signs List 30 main body 31 input means 39 printer 38 display 40 external recording means 101 two-character chain detector 102 three-character chain detector 103 two-character chain memory 104 three-character chain memory 111 two-character chain detector 112 three-character chain detector 113 comparator 114 control unit 301 character string converter 302 two-character chain detector 303 two-character chain memory 304 character string converter 305 two-character chain detector 306 comparator 307 control unit 501 character string converter 502 two-character chain detector 503 two-character Chain memory 504 Character string converter 505 Two-character chain detector 506 Comparator 507 Control unit 901 Special character detector 902 Two-character chain detector 903 Special character chain detector 904 Two-character chain memory 911 Special character detector 912 Two-character chain Detector 913 Special character chain detection Output unit 914 Comparator 915 Control unit 1101 Two-character chain detector 1102 Three-character chain detector 1103 Two-character chain memory 1104 Three-character chain memory 1111 Two-character chain detector 1112 Three-character chain detector 1113 Comparator 1114 Control unit 1301 2 Character chain detector 1302 Three character chain detector 1303 Special two character chain generator 1304 Two character chain memory 1311 Two character chain detector 1312 Three character chain detector 1313 Special two character chain generator 1314 Comparator 1315 Control unit 1501 Registered characters Sequence 1502 Two-character chain 1503 Two-character chain 1504 Two-character chain 1505 Two-character chain 1506 Two-character chain 1507 Two-character chain 1508 Two-character chain 1509 Two-character chain 1510 Two-character chain 1511 Two-character chain 1512 Two-character chain 15 3 Two-character chain 1514 Two-character chain 1515 Two-character chain 1516 Occurrence overlap 1517 Occurrence overlap 1601 Special character detector 1602 Two-character chain detector 1603 Special character chain detector 1604 Occurrence overlap memory 1605 Special character chain sorter 1606 Two characters Chain memory 1607 Special character detector 1608 Two-character chain detector 1609 Special character chain detector 1610 Comparator 1611 Controller 1612 Appearance duplicate counter memory 1701 Document number 1702 Number of first character appearances 1703 Number of second character appearances 1704 First Number of occurrences of special characters 1705 Number specified for second character 1706 Number specified for first character 1707 Number of occurrences of special character for second character 1708 Registered character string 1709 Two-character chain 1710 Two-character chain 1711 Two-character chain 17 12 Two-character chain 1713 Two-character chain 1714 Two-character chain 1717 Two-character chain 1716 Two-character chain 1717 Two-character chain 1718 Two-character chain 1719 Two-character chain 1720 Two-character chain 1721 Character-chain data 1722 Character-chain data 1723 Character-chain data 1724 characters Chained data 1725 Character chained data 1726 Character chained data 1727 Character chained data 1728 Character chained data 1729 Character chained data 1730 Character chained data 1731 Character chained data 1732 Search character string 1733 Two character chain 1734 Two character chain 1735 Two character chain 1736 Two character chain 1737 Special character appearance counter 1738 Special character appearance counter 1801 to 1814 Step 1901 to 1912 Step 2001 Two-character chain detector 20 2 Appearance frequency calculator 2003 Special character chain detector 2004 Two-character chain memory 2005 Two-character chain detector 2006 Comparator 2007 Special character appearance counter memory 2008 Controller 2101 Special character detector 2102 Two-character chain detector 2103 Special two-character chain Detector 2104 Two-character chain memory 2105 Special character detector 2106 Two-character chain detector 2107 Special two-character chain detector 2108 Comparator 2109 Controller 2201 Registered character string 2202 Two-character chain 2203 Special two-character chain 2204 Two-character chain 2205 Search Character string 2206 Two-character chain 2207 Special two-character chain 2208 Two-character chain 2301 Two-character chain position detector 2302 Three-character chain position detector 2303 Two-character chain position memory 2304 Three-character chain position memory 2311 2 Character chain detector 2312 3 character chain detector 2313 comparator 2314 control unit 2501 character string converter 2502 2 character chain position detector 2503 2 character chain position memory 2504 character string converter 2505 2 character chain detector 2506 comparator 2507 control Unit 2701 character string converter 2702 two-character chain position detector 2703 two-character chain position memory 2704 character string converter 2705 two-character chain detector 2706 comparator 2707 control unit 2901 three-character chain detector 2902 two-character chain position detector 2903 Special two-character chain generator 2904 Two-character chain position memory 2905 Special two-character chain position memory 2911 Three-character chain detector 2912 Two-character chain detector 2913 Special two-character chain generator 2914 Comparator 2915 Control unit 3101 Special character detector 3 102 Two-character chain encoder 3103 Special two-character chain encoder 3104 Character chain combination determiner 3105 Two-character chain memory 3106 Special character detector 3107 Two-character chain detector 3108 Special two-character chain detector 3109 Character chain combination determiner 3110 Comparison Unit 3111 Judgment unit 3201 Registered character string 3202 Two-character chain 3203 Two-character chain 3204 Two-character chain 3205 Two-character chain 3206 Special two-character chain 3207 Special two-character chain 3208 Special two-character chain 3209 Special two-character chain 3210 Special two-character chain 3211 Special two-character chain 3212 Special two-character chain 3213 Special two-character chain 3214 Special two-character chain 3215 Document number 3216 Registered character position 3217 Search character string 3218 Two-character chain 3219 Special two-character chain 3220 Special Two-character chain 3221 Special two-character chain 3222 Two-character chain 3223 Search character string 3224 Two-character chain 3225 Special two-character chain 3226 Special two-character chain 3227 Special two-character chain 3228 Two-character chain 3301 Special character detector 3302 Character chain combination determiner 3303 two-character chain encoder 3304 special two-character chain encoder 3305 special two-character chain encoder 3306 two-character chain memory 3307 character chain combination determiner 3308 two-character chain detector 3309 special two-character chain detector 3310 comparator 3311 determination unit 3401 Registered character string 3402 Registered character string 3403 Two-character chain 3404 Special two-character chain 3405 Two-character chain 3406 Two-character chain 3407 Two-character chain 3408 Two-character chain 3409 Registered character position 3410 Character position 3411 Character chain information 3412 Document number 3413 Two character chain 3414 Character position 3415 Special two character chain 3416 Character position 3417 Special two character chain flag 3418 Character chain information 3419 Two character chain 3420 Special two character chain 3421 Special two character chain flag 3422 Character chain information 3423 Special two character chain flag 3424 Character chain information 3425 Two character chain 3426 Special two character chain 3427 Special two character chain flag 3428 Character chain information 3429 Special two character chain flag 3501 Search character string 3502 Search character string 3503 Two character chain 3504 Special 2 Character chain 3505 Search character string 3506 Search character string 3507 Special two-character chain 3508 Search character string 3509 Search character string 3510 Special two-character chain 3511 Search character string 3512 Search character 3513 two-character chain 3514 two-character chain 3515 special two-character chain

 ──────────────────────────────────────────────────の Continued on the front page (72) Tomoko Fujita, Inventor 1006 Kazuma Kadoma, Osaka Prefecture Matsushita Electric Industrial Co., Ltd. (72) Inventor Yasuhiro Shirasaki 1006 Kadoma Kadoma, Kadoma City, Osaka Matsushita Electric Industrial Co.

Claims (30)

[Claims] 1. A computer-readable recording medium in which search data used for full-text search is recorded, wherein the search data includes all characters including characters other than a special character specified in advance for a search target character string. First data in which a character chain is detected, and for each two-character chain, first data that records the number of occurrences of the first character and the second character that form the two-character chain in the search target character string, and a special character specified in advance 2 other than special characters in which is inserted
Detects all character chains composed of characters, and stores, for each character chain, second data in which the number of occurrences of the first character and the second character constituting the character chain in the search target character string is recorded as a first data. A computer-readable recording medium on which full-text search data is recorded, wherein the data and the second data are recorded separately. 2. A computer-readable recording medium on which search data used for full-text search according to claim 1 is recorded, and all two-character sequences consisting of characters other than special characters specified in advance are detected from a search character string. A first character chain detecting means for detecting all character chains consisting of two characters other than the special character in which a predetermined special character is inserted from a search character string; The two-character chain detected by the first character-chain detecting means is searched from the first data recorded on the recording medium, and the character chain detected by the second character-chain detecting means is detected. A character string collating device comprising: comparing means for judging the presence or absence of a continuation of a character chain as a search character string by comparing the number of appearances corresponding to the character chain. 3. A computer-readable recording medium storing search data used for full-text search, wherein the search data is not a search target according to a character adjacent to a special character designated in advance in a search target character string. Character strings, and detects all two-character chains in the converted character string including the characters not to be searched.
A computer-readable recording medium on which full-text search data is recorded, wherein the number of appearances of a first character and a second character constituting a two-character chain in a character string to be searched is recorded as a set for each character chain. 4. A computer-readable recording medium on which search data used for full-text search according to claim 3 is recorded, and a special character specified in a search character string is applied to data recorded on said recording medium. Subject to the same rules,
A character string conversion unit for converting a character string that is not a search target based on adjacent characters; and detecting a two-character chain including a character that is not a search target for the character string converted by the character string conversion unit. A two-character chain detecting unit, and a two-character chain detected by the two-character chain detecting unit is detected from the recording medium, and the presence or absence of the continuation of the character chain as a search character string is determined by comparing the corresponding number of appearances. A character string collating device, comprising: 5. A computer-readable recording medium on which search data used for full-text search is recorded, wherein the search data is not a search target according to a character adjacent to a special character specified in advance in a search target character string. Converted to two characters, and detected all two-character chains in the converted character string, including the characters that are not to be searched.
For each two-character chain, the first and second characters that make up the two-character chain
A computer-readable recording medium on which full-text search data is recorded, wherein the number of appearances of a character in a character string to be searched is recorded as a set. 6. A computer-readable recording medium on which search data used for full-text search according to claim 5 is recorded, and a special character specified in a search character string is applied to data recorded on said recording medium. Subject to the same rules,
A character string conversion means for converting into two characters that are not to be searched based on adjacent characters; and a two-character chain including two characters that are not to be searched for the character string converted by the character string conversion means. A two-character chain detecting means for detecting the two-character chain detected by the two-character chain detecting means from the recording medium, and comparing the corresponding number of appearances to determine whether or not the character string as a search character string is continuous; A character string collating device, comprising: comparing means for judging a character string. 7. A computer-readable recording medium recording search data used for full-text search, wherein the search data detects a two-character chain for all characters in a search target character string, For each of the first and second characters constituting a character chain other than the special character specified in advance, which constitutes a two-character chain for each chain, the number of appearances of the characters other than the special character specified in advance is specified in advance. In the case of special characters, all three-character sequences consisting of three characters with a special character specified in advance between the third data recorded as a set of certain numerical values and the search target character string are detected. And
For each three-character chain, the first and third characters that make up the three-character chain
A computer-readable recording of full-text search data, characterized in that the fourth data recorded as a set of the number of appearances of the character in the search target character string and the third data and the fourth data are recorded separately. Recording medium. 8. A computer-readable recording medium on which search data used for full-text search according to claim 7 is recorded, and all two-character sequences consisting of characters other than special characters specified in advance are detected from a search character string. A first character chain detecting means for detecting, from the search character string, all character chains consisting of three characters into which a special character designated in advance has been inserted; and a first character chain detecting means. The two-character chain detected by the detecting means is searched from the first data recorded on the recording medium, the character chain detected by the third character chain detecting means is detected, and the character chain corresponding to the detected character chain is detected. A character string collating device comprising: comparing means for judging the presence or absence of continuation of a character chain as a search character string by comparing the number of appearances. 9. A computer-readable recording medium on which search data used for full-text search is recorded, wherein the search data includes all two-byte characters other than special characters specified in advance for a character string to be searched. Fifth data in which a character chain is detected and the number of appearances of the first character and the second character constituting the two-character chain in the search target character string are recorded as a set, and for the search target character string, Detects all three-character chains consisting of three characters with special characters specified in advance,
For each three-letter chain, a sixth set in which a set of the number of appearances of the first character and the value 0 constituting the three-letter chain and a value 0 and a set of the number of occurrences of the value 0 and the third character as a set of two sets are recorded. A computer-readable recording medium on which full-text search data is recorded, wherein the data, the fifth data, and the sixth data are recorded separately. 10. A computer-readable recording medium on which search data used for full-text search according to claim 9 is recorded, and all two-character sequences consisting of characters other than special characters specified in advance are detected from a search character string. A first character chain detecting means for detecting, from the search character string, all character chains consisting of three characters into which a special character designated in advance has been inserted; and a first character chain detecting means. The two-character chain detected by the detecting means is searched from the first data recorded on the recording medium, the character chain detected by the third character chain detecting means is detected, and the character chain corresponding to the detected character chain is detected. A character string collating device comprising: comparing means for judging the presence or absence of continuation of a character chain as a search character string by comparing the number of appearances. 11. A computer-readable recording medium in which search data used for full-text search is recorded, wherein the search data includes all two-byte characters other than special characters specified in advance for a character string to be searched. Seventh data in which a character chain is detected, and the number of occurrences of the first character and the second character constituting the two-character chain in the character string to be searched for each two-character chain are recorded as a set. Detects all three-character chains consisting of three characters with special characters specified in advance,
For each three-character chain, the special character of the second character constituting the three-character chain is converted into the same character as the third character, and the number of occurrences of the second character is set to the same value as the number of occurrences of the third character, and then the first character A two-character chain consisting of a second character, a second character, and a third character is generated, and for each two-character chain, the number of appearances of the first character and the second character constituting the two-character chain in the search target character string is set. A computer-readable recording medium on which full-text search data is recorded, characterized in that the eighth data recorded as the first data and the seventh data and the eighth data are recorded separately. 12. A computer-readable recording medium on which search data used for full-text search according to claim 11 is recorded, and all two-character sequences consisting of characters other than special characters specified in advance are detected from a search character string. A first character chain detecting means for detecting all three-character chains consisting of three characters into which a predetermined special character is inserted, from a search character string; and a first character chain detecting means. The two-character chain detected by the chain detecting means is searched from the first data recorded on the recording medium, and the special character detected by the fourth character chain detecting means is converted to generate a two-character chain. A character string collating apparatus, comprising: comparing means for comparing the number of appearances corresponding to each two-character chain to determine whether or not there is a continuation of character chains as a search character string. 13. A computer-readable storage medium storing search data used for full-text search, wherein the search data detects all two-character sequences in a search target character string, If the two-character chain is a character chain other than the special character specified in advance, the characters other than the special character specified in advance for the first character and the second character are counted as the number of appearances. In the case where the ninth data that records the set and the two-character chain is a character chain that includes a special character specified in advance, the number of appearances of the first or second character corresponding to the special character is Remainder divided by less than the maximum value of the number of appearances specified in advance, or the maximum value if the remainder is 0, or the maximum value and the remainder, or the second and subsequent times if the first occurrence number is less than the maximum value Maximum value of When the lower value has a value such that the first value and the order are unique, the number of appearances of a character that is not a special character is stored as a set, and when the first character is a special character, this data set is: A computer-readable storage medium storing full-text search data, wherein tenth data sorted for each second character type, and ninth and tenth data are stored separately. 14. A computer-readable storage medium storing search data used for full-text search according to claim 13, and a fifth character for detecting all two-character sequences that do not include special characters from a search character string. A chain detecting means, a sixth character chain detecting means for detecting all character chains including special characters from the search character string, and a second character chain detecting means for detecting the search character string by the fifth character chain detecting means.
In the case where the character string is composed of character chains, a comparison unit that determines whether or not there is a continuation of character chains as a search character string by comparing the number of appearances corresponding to the detected character chain; 2 searched by character chain detection means
In the case of a character chain, a comparison means is provided for judging whether or not there is a continuation of the character chain as the search character string by comparing the number of occurrences of the detected character chain and the number of times of occurrence of the special character. A character string collating device characterized by the following. 15. A computer-readable storage medium on which search data used for full-text search is recorded, wherein the search data detects all two-character sequences in a search target character string, and outputs a document for each two-character sequence. Numbers, the number of appearances for each character type of the two-character chain, or character chain data composed of a set of arbitrary values. If the character string data does not include a special character specified in advance, the number of appearances of the first character is determined. If the size for storing the number of appearances of the second character is equal and includes a special character specified in advance, the size for storing the number of appearances corresponding to the special character is stored as an arbitrary value corresponding to a character not including the special character. In the first character chain data configured to be larger than the size, if the first character chain data includes a special character string specified in advance as the first character, the second character is a finger. Is stored, and the second character chain data is configured so that the first character of the next continuous character chain data is equal to the value specified by the second character of the previous character chain data. A computer-readable storage medium on which full-text search data is recorded. 16. A computer-readable storage medium storing search data used for full-text search according to claim 15, and a fifth character for detecting all two-character sequences that do not include special characters from a search character string. A chain detecting means, and if the two-character chain does not include a special character, the character chain data corresponding to the continuous character chain detected by the fifth character chain detecting means is compared with the character chain data of the detected character chain data. Comparing means for comparing the number of appearances of the two characters with the number of appearances of the first character of the character chain data of the character chain following the character chain to determine whether or not there is continuation of the character chain as a search character string; If the character chain includes special characters, the character chain data corresponding to the continuous character chain searched by the character chain detection unit is compared with the number of occurrences of the character in the same manner as the comparison unit. Means for storing the number of occurrences of a specified special character string at the time of comparison, comparing the character strings on the basis that the number of occurrences does not overlap except for a continuous character chain, and determining whether or not there is a continuation of the character chain as a search character string A character string collating device comprising: 17. A computer-readable storage medium storing search data used for full-text search, wherein the search data detects all two-character chains that do not include special characters in a search target character string, The first character and the second character constituting a character chain other than the special character designated in advance, which constitutes every two character chains.
Regarding characters, characters other than the special characters specified in advance are recorded as a set of the eleventh data of the number of appearances of the special characters. Second data recorded as a set of the number of appearances of the first character type in the two-character chain preceding the two-character chain and the number of occurrences of the first character type in the two-character chain following the special character, or before the special character , The twelfth data, which is a combination of the number of appearances of the first character type of the two-character chain and the number of appearances of the character immediately after the special character, and the eleventh data and the twelfth data are recorded separately. A computer-readable recording medium on which full-text search data is recorded. 18. A computer-readable storage medium on which search data used for full-text search according to claim 17 is recorded, and a two-character chain consisting of characters other than special characters designated in advance from a search character string. A first character chain search means for detecting, a first character of the two-character chain before the special character, and a second character chain after the special character from the special character string with respect to the two-character chain before and after the special character specified in advance. Or a character chain that combines the first character of the two-character chain preceding the special character and the character that immediately follows the special character. A seventh character chain detecting means to be detected, and a two-character chain detected by the first character chain detecting means are searched from the first data or the twelfth data recorded on the storage medium. Search from 11 data When the character string is searched, the character chain detected by the seventh character chain detecting means is searched. When the character string is searched from the twelfth data, the character chain detected by the first character chain detecting means is searched. A character string collating device comprising: a comparing unit that determines whether or not there is a continuation of a character chain as a search character string by comparing the number of appearances corresponding to the character chain. 19. A computer-readable recording medium recording search data used for full-text search, wherein the search data includes a special character specified in advance, and counts only occurrences of characters other than the special character specified in advance. Then, in the character string that is the appearance position of each character, a chain of all two characters consisting of characters other than the special character specified in advance is detected from the search target character string, and a two-character chain is detected for each two-character chain. And the thirteenth data in which the appearance position of the first character forming the character string is recorded as the appearance position in the search target character string, and all the character chains in which the special characters specified in advance have been inserted are detected. Fourteenth data, in which the appearance position of the first character constituting the character chain is recorded as the appearance position in the search target character string, is recorded separately from the thirteenth data and the fourteenth data Wherein, the computer-readable recording medium recording a full-text search data. 20. A computer-readable recording medium on which search data used for full-text search according to claim 19 is recorded, and all two-character sequences consisting of characters other than special characters specified in advance are detected from a search character string. A first character chain detecting means for detecting all character chains in which a predetermined special character is inserted from a search character string; and a first character chain detecting means for detecting the character string. Character chain
Searching from the thirteenth data recorded on the recording medium,
The character chain detected by the eighth character chain detecting means is
Searching from the fourteenth data recorded on the recording medium,
A character string collating device comprising: comparing means for judging the presence or absence of continuation of character chains as a search character string by comparing appearance positions corresponding to character chains. 21. A computer-readable recording medium on which search data used for full-text search is recorded, wherein the search data is a character string containing a special character specified in advance, and a special character string of a character string to be searched is specified. Characters are converted into characters that are not to be searched according to adjacent characters, and all two-character sequences including the characters that are not to be searched are detected from the converted character string. A computer-readable recording medium on which full-text search data is recorded, wherein an appearance position of a first character forming a character chain is recorded as an appearance position in a search target character string. 22. A computer-readable recording medium on which search data used for full-text search according to claim 21 is recorded, and a special character specified in a search character string is applied to data recorded on said recording medium. Subject to the same rules,
A character string conversion unit for converting a character string that is not a search target based on adjacent characters; and detecting a two-character chain including a character that is not a search target for the character string converted by the character string conversion unit. A two-character chain detecting unit, and a two-character chain detected by the two-character chain detecting unit is detected from the recording medium, and the presence or absence of the continuation of the character chain as a search character string is determined by comparing corresponding appearance positions. A character string collating device, comprising: 23. A computer-readable recording medium on which search data used for full-text search is recorded, wherein the search data is a character string including a special character specified in advance, and a special character string of a character string to be searched is specified. Characters are converted into two characters that are not to be searched according to adjacent characters, and for the converted character string, all two-character chains including the characters that are not to be searched are detected. 2
A computer-readable recording medium on which full-text search data is recorded, wherein an appearance position of a first character or a second character constituting a character chain is recorded as an appearance position in a search target character string. 24. A computer-readable recording medium on which search data used for full-text search according to claim 23 is recorded, and a special character specified in a search character string is applied to data recorded on said recording medium. Subject to the same rules,
A character string conversion means for converting into two characters that are not to be searched based on adjacent characters; and a two-character chain including two characters that are not to be searched for the character string converted by the character string conversion means. A two-character chain detecting means for detecting, and a two-character chain detected by the two-character chain detecting means is detected from the recording medium, and a corresponding appearance position is compared to determine whether or not the character chain as a search character string is continuous. A character string collating device, comprising: comparing means for judging a character string. 25. A computer-readable recording medium on which search data used for full-text search is recorded, wherein the search data is a character string including a special character specified in advance, and a character string specified in advance for a search target character string. In a character string consisting of characters other than special characters, all two-character
The first or second character that constitutes a two-character chain for each character chain
For the data in which the character appearance position is the appearance position of the search target character string, and for the search target character string, all three-character chains consisting of three characters with special characters specified in advance are detected. For each character chain, the special character of the second character constituting the three-character chain is converted into the same character as the third character, and the two-character chain consisting of the first character and the second character is detected to form the two-character chain. The fifteenth data which records data in which the appearance position of the first character or the second character to be searched is the appearance position of the search target character string, and the two-character sequence consisting of the second character and the third character of the three-character sequence After detecting the occurrence position of the first character in the two-character chain as the same value as the appearance position of the second character, the appearance position of the first character or the second character is determined as the appearance position of the character chain in the search target character string. The recorded 16th data and the 15th data Data and wherein the the first 16 data are recorded in distinction, a computer readable recording medium recording a full-text search data. 26. A computer-readable recording medium on which search data used for full-text search according to claim 25 is recorded, and all two-character sequences consisting of characters other than special characters specified in advance are detected from a search character string. A first character chain detecting means for detecting all three-character chains consisting of three characters into which a special character specified in advance has been inserted, from a search character string; and a first character chain detecting means. The two-character chain detected by the chain detecting means is searched from the fifteenth data recorded on the recording medium, and the special character detected by the tenth character chain detecting means is converted to generate a two-character chain. A character string collating apparatus, comprising: comparing means for comparing appearance positions corresponding to each two-character chain to determine whether or not there is a continuation of character chains as a search character string. 27. A computer-readable storage medium storing search data used for full-text search, wherein the search data detects a two-character chain and a character position in a search target character string, A set of a document number, a two-character chain, and a character position composed of the search target character string is configured as two-character chain information, and the character position of the two-character chain information is specified in advance based on the beginning of the search target character string. A storage medium storing character chain information for each first character of a character chain, excluding special character positions and numbering in ascending order, wherein the first character is a two-character chain that does not include special characters. The seventeenth data recorded as a set of the first character at the character position excluding the special character and the character position and the document number excluding the special character, and the character immediately before the special character and the special character Two characters The 18th data consisting of a chain, an arbitrary fixed value defined by the character type of the special character, and a document number, a two-character chain combining the special character and the character immediately following the special character, and the special character A two-character chain combining the nineteenth data consisting of the character position of the second character at the excluded character position and the document number and the characters immediately before and after the special character, and the character position at the character position excluding the special character 20th data composed of the character position of the first character and the document number; and 17th data, 18th data, 19th data, and 20th data. It is sorted and stored as character chain information. For the seventeenth data, the eighteenth data, and the nineteenth data, if the first character of the two character chains is the same and the second character is a special character, 17
A computer-readable storage medium on which full-text search data is recorded, wherein the eighteenth data is stored immediately after the data. 28. A computer-readable storage medium storing search data used for full-text search according to claim 27, and detecting all two-character sequences excluding a special character and characters before and after the special character from a search character string. An eleventh character chain detecting means, a two-character chain consisting of a character immediately before and after a special character, a two-character chain consisting of a character immediately before a special character and a special character, a special character and a special character from a search character string A twelfth character chain detecting means for detecting a two-character chain consisting of the character immediately after the character string, a seventeenth data corresponding to the two-character chain detected by the first character chain detecting means, and a twelfth character chain detecting means. 20th data corresponding to the character chain detected in
Comparing means for determining the presence or absence of continuation as a search character string by comparing the character position difference between the two data from the seventeenth data with the document number; and the eighteenth data immediately after the twentieth data. Comparing means for determining the presence or absence of a continuation as a search character string including a special character by being continued. 29. A computer-readable storage medium storing search data used for full-text search, wherein the search data detects a two-character chain and a character position with respect to a character string to be searched. A storage medium comprising a set of a document number composed of a search target character string, a two-character chain, and a character position as two-character chain information, and storing character chain information for each first character of the character chain, The character position of the two-character chain information is numbered in ascending or descending order excluding the position of the special character specified in advance with respect to the beginning of the search target character string. The 21st data recorded as a set of one character and the second character, the character position of the first character, the document number, and the character chain information including the special character are the same as the character just before the special character. , Just before special characters A character chain combining a character and a character immediately following a special character, a set of a character position immediately before the special character and a document number, and the first character and the second character of the character chain of the character chain information are special characters. When the first character or the second character of the character chain of the character chain information of the character chain information that does not include the special character is matched, the 22nd character string is separately recorded after or before the character chain information that does not include the special character. Data and character chain information including special characters, for the character immediately following the special character, the character chain combining the character immediately following the special character and the character that follows that character, the character position immediately following the special character And a document number, and when the first character of the character chain of the character chain information matches the first character of the two-character chain when no special character is included, the character chain information without the special character is included. After or before A second 23 data configured to be recorded in the number, character chain information including special characters, character chain that combines the character after the two previous characters and special characters special characters,
The twenty-fourth data consisting of the character position and the document number two places before the special character, and the twenty-first data, the twenty-second data, the twenty-third data, and the twenty-fourth data are stored separately. Computer-readable storage medium that records full-text search data. 30. A computer-readable storage medium storing the search data used in the full-text search according to claim 29, and all the characters other than characters before and after the special character from the search character string.
A thirteenth character chain detecting means for detecting a character chain; and in the case of a search character string sandwiching a special character from the search character string, a character immediately before and after the special character is detected as a character chain; The second character in the character chain is marked as the character immediately following the special character. If the beginning of the search character string is a special character, the character immediately after the special character and the next character are detected as a character chain, and the character chain is detected. Is marked as the character immediately after the special character, and if the special character appears after the third character from the beginning of the search string, the character two characters before the special character and the character immediately after the special character Detected as a character chain and the second of the character chain
The character is marked as the character immediately following the special character, the character immediately following the special character and the next character are detected as a character chain, and the first character of the character chain is marked as the character immediately following the special character. In the case where the fourteenth character chain detecting means and the search character string are composed of two character chains detected by the thirteenth character chain detecting means, a comparison between the character position corresponding to the detected character chain and the document number is performed. The comparison means for judging the presence or absence of continuation of the character chain as the search character string, and the twenty-first character string when the search character string is composed of the two character chain searched by the fourteenth character chain detection means. Comparing means for determining whether or not the data matches the character chain information of the twenty-fourth data based on the character chain and the document number to determine whether or not there is a continuation of character chains as a search character string.