JPH06301730A - Data retrieving device - Google Patents

Abstract

PURPOSE:To provide the data retrieving device which can execute the processing of a multi-type record tra. CONSTITUTION:A designating means 2 receives an input from a user, and sends the information required for the processing to a generating means 3. The generating means 3 receives the information required for the processing, sent from the designating means 2, and thereafter, reads in retrieval object data having an identifier designated in its information from a store means 1, executes the processing of the read-in retrieval object data, and generates new retrieval object data. The new retrieval object data is displayed by a display means 4, and a user can look at a result of retrieval. The retrieval object data can be constituted of a multi-type record train having a sequential relation consisting of plural kids of records such as logical information of a document. The processing executed by generating means 3 can be executed for the multi-type record train.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data retrieval device for a user to retrieve desired information from a record string.

[0002]

2. Description of the Related Art There is a relational database as a retrieval device for a user to retrieve desired information from accumulated data. The relational database is a relational model (Communication of t) based on the first normal form.
he ACM, Vol. 13, No. 6, June 1
970, Codd, E .; F. , "A Relation
al Model of Data for Larg
e Shared Data Banks. ", Pp. 3
77-387), any field in any record in the relational database has no internal structure, or the relational database cannot access the internal structure.

To solve this problem, Damam,
P. Kuesect, K .; , Anderson,
F. Blanken, H .; Erbe, R .; , Gue
nuer, J .; Lum, V .; , Pistor,
P. , And Walch, G .; , "A DBMS P
rototype to Support Extension
ded NF ² Relations: An Inte
grated Viewon Flat Tables
and Hierarchies ”, SIGMOD,
1986, pp. As in 356-367, an extended relational database for the extended relation that allows a hierarchical structure in which the field value of the record in the database is a record group without the restriction of the first normal form has been considered. In extended relational algebra for manipulating extended relational databases, Roth, M. et al. A. , Korth, H .; F. , A
nd Silberschats, A .; , "Exten
ded Algebra and Calculus
for Nested Relational Dat
As described in “Abase”, unions and products provided in conventional relational algebra.
section), difference (set difference)
e), Cartesian product
t), projection, selection (sel)
In addition to operations such as "action", it provides a function of hierarchical or non-hierarchical data.

NST (Ne is used to apply the concept of expansion relation to a database for documents.
sted Sequences of Tuples)
And its algebra. This NST is, for example, ACM T
transaction onOffice Info
mation Systems, Vol. 7, No.
4, April 1989, Guting, R .; H, Z
icari, R .; , AND Choy, D.M. M, "An
algebra for structured o
"fice documents".
NST is a data model for expressing the logical structure of a structured document. The NST brings into the expanded relationship the characteristic of the document that it consists of a sequence of components. The extended relation and its algebra deal with a set of records, whereas N
ST and its algebra deal with columns of records. NST
And its algebra, the sum (interion), the difference (set difference), and the orthogonal product (car) that are provided in the conventional extended relation.
tesian product), projection (projec)
functions of selection, selection, layering, and non-layering are changed to those that handle the columns of records. A document generally has an order relation, and when handling a document, it is necessary to perform processing in consideration of this order relation. Therefore, the record set handled in the database is treated as a sequence of records having an order relation. Further, the functions provided in the conventional extended relationship also take this order relationship into consideration.

However, NST is not sufficient to represent the logical structure of a document. Relational database, extended relations,
As it handled a set of records of the same kind, NS
T treats the document as a sequence of records of the same type. On the other hand, an actual document is a sequence of different types of components. For example, it is common for a document to be a sequence of different types of data such as figures, tables, paragraphs. Also, the document
It is also a sequence of data that has constituent elements such as chapters and data that does not have constituent elements such as paragraphs. NST cannot express the characteristics of a document that is a string of data of different types.

[0006]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a data search device capable of processing a multi-type record string.

[0007]

According to a first aspect of the present invention, in a data search apparatus, a plurality of types of records are used in a data search apparatus that retrieves desired information from a record string having an order relation. Storage means for storing a record string that can be configured by the storage means, designating means capable of inputting a processing request for the record series stored in the storage means, and storing in the storage means based on the processing request input from the designating means. The present invention is characterized by having a generation means for performing processing on the existing record string to generate a new record string.

According to a second aspect of the present invention, in the data search apparatus according to the first aspect, the type of the record and the condition for the record of the type are designated by the designating unit, and the generating unit stores the storage. A new second record string is generated by removing from the first record string stored in the means a record that is of the type of the record specified by the specifying means and does not satisfy the condition. To do.

According to a third aspect of the present invention, in the data search apparatus according to the first aspect, the type of the record and the condition for the record of the type are designated by the designating unit, and the generating unit stores the storage. A record that is the type of the record designated by the designation unit and does not satisfy the condition, from the first record string stored in the unit;
It is characterized in that a second record string is generated by removing records not designated by the designating means.

According to the invention of claim 4, in the data search apparatus according to claim 1, the generating means concatenates the second record string after the first record string stored in the storage means. It is characterized in that a new third record string is created.

According to a fifth aspect of the present invention, in the data search apparatus according to the first aspect, the generating means stores the first record string stored in the storing means in the storing means. It is characterized in that a new third record string is generated by removing the records that do not exist in the existing second record string.

According to a sixth aspect of the present invention, in the data search apparatus according to the first aspect, the generating means stores the first record string stored in the storing means in the storing means. It is characterized in that a new third record string is generated by removing the records existing in the existing second record string.

According to a seventh aspect of the present invention, in the data search apparatus according to the first aspect, the generating means includes a field of a record included in the first record string stored in the storage means, and The records arranged in this order with the fields of the records included in the second record string stored in the storage means are included in the second record string in the order of the records included in the first record string. It is characterized in that a new third record string created for all combinations in the order of the records to be created is generated.

According to an eighth aspect of the present invention, in the data search apparatus according to the first aspect, the designation means designates a record type and a field arrangement order of the records of the type, and the generation means is , A new second record string obtained by changing the order of fields of the records of the type designated by the designating unit of the first record sequence stored in the storage unit, to the sequence order designated by the designating unit. It is characterized by generating.

According to a ninth aspect of the present invention, in the data search apparatus according to the first aspect, the designation means designates a record type and a field arrangement order of the records of the type, and the generation means is Changing the arrangement order of the fields of the record of the type designated by the designating means of the first record string stored in the storage means, and further changing the order of the fields designated by the designating means. A new second that removes unspecified types of records
It is characterized by generating a record string of.

According to a tenth aspect of the present invention, in the data search apparatus according to the first aspect, the type of the record and a part of the field group of the record of the type are designated by the designating unit, and the generating unit is provided. Is a record of the type designated by the designating means of the first record string stored in the storing means as a new record consisting of only the fields designated by the designating means among the fields in the record. It is characterized in that a new replaced second record string is generated.

According to an eleventh aspect of the present invention, in the data search apparatus according to the first aspect, the type of record and a part of a field group of the record of the type are designated by the designating unit, and the generating unit is provided. Is a new record consisting of only the fields designated by the designating means among the fields in the record of the type designated by the designating means of the first record string stored in the storing means. It is characterized in that a new second record string is generated by replacing and further removing a record of a type not designated by the designation means.

According to a twelfth aspect of the invention, in the data search apparatus according to the first aspect, the type of the record and a part of the field group of the record of the type are designated by the designating means, and the generating means is provided. Is a partial sequence of the first record sequence stored in the storage unit, which is of the type designated by the designating unit and has the same field value of the field group designated by the designating unit. A new second that replaces a new record created by layering columns
It is characterized by generating a record string of.

According to a thirteenth aspect of the present invention, in the data search apparatus according to the first aspect, the type of record and a part of the field group of the record of the type are designated by the designating unit, and the generating unit is provided. Is a subsequence of the records of the first record sequence stored in the storage means, which is of the type designated by the designating means and has the same field value of the field group designated by the designating means. , A new second record string is created by replacing the partial string with a new record created by hierarchizing, and further removing a record of a type not specified by the specifying means.

According to the invention described in claim 14, in claim 12,
Alternatively, in the data search device described in the paragraph 13, the new record created by hierarchizing the partial sequence is a record including a field group designated by the designating unit and a field having the partial sequence as a value. It is a feature.

According to the invention of claim 15, in claim 12,
Alternatively, in the data search device described in the paragraph 13, a new record created by hierarchizing the partial sequence includes a field group designated by the designating means and each record in the partial sequence among the field groups of the record. The record is composed of a field whose value is the value replaced by the record not specified by the specifying means.

According to a sixteenth aspect of the present invention, in the data search apparatus according to the first aspect, the type of record and a part of the field group of the record of the type are designated by the designating unit, and the generating unit is provided. Is one or more records in the first record string stored in the storage means, of the type specified by the specification means, and having the same field value of the field group specified by the specification means. Records are sequentially extracted and concatenated, and a new record created by hierarchizing this record string and a record in the first record string that is not specified by the specifying means are It is characterized in that a new second record string in which the records in the record string are connected in order is generated.

According to a seventeenth aspect of the present invention, in the data search apparatus according to the first aspect, the designation unit designates a record type and a part of a field group of the record of the type, and the generation unit. Is one or more records in the first record string stored in the storage means, of the type specified by the specification means, and having the same field value of the field group specified by the specification means. Is sequentially extracted and concatenated, and a new record created by layering this record sequence is connected to the records in the first record sequence to generate a new second record sequence. Is.

According to the invention described in claim 18, claim 16 is provided.
Alternatively, in the data search device described in the paragraph 17, the new record is a record including a field group designated by the designating unit and a field having a value as a record string obtained by concatenating the one or more records in the order of extraction. It is characterized by being.

In the invention described in claim 19, claim 16 is provided.
Alternatively, in the data search device described in the paragraph 17, the new record is designated by the designating unit from each field of the field group designated by the designating unit and each record of a record string in which the one or more records are connected in the order of extraction. It is characterized in that it is a record consisting of a field whose value is a record string obtained by removing the field.

According to a twentieth aspect of the invention, in the data retrieving apparatus according to the first aspect, a field, which is a record string consisting of records of one or more types of record and field value, is specified by the specifying means. Is specified, the generating means specifies, from the first record string stored in the storing means, a record of the type specified by the specifying means, from the specifying means in the field of the record. A new second record string is generated by replacing each record of the record string that is the value of the field with a record string created by adding to a field group not specified by the specifying means in the record. To do.

According to the twenty-first aspect of the present invention, in the data search apparatus according to the first aspect, a field which is a record string including a record type and a record having one or more field values from the specifying means. Is specified, the generating means specifies, from the first record string stored in the storing means, a record of the type specified by the specifying means, from the specifying means in the field of the record. Replace each record of the record string that is the value of the field with a record string created by adding to the field group not specified by the specifying means in the record, and further, record of a type not specified by the specifying means It is characterized in that a new second record string from which is deleted is generated.

According to the twenty-second aspect of the present invention, in the data search apparatus according to the first aspect, the type of record and a creating method for creating new data from the record of the type are specified by the specifying means. The generating means stores data created from the record in the first record string stored in the storing means in the record of the type specified by the specifying means according to the creating method specified by the specifying means. It is characterized in that a new second record string to which a field having a field value of is added is generated.

In the invention described in Item 23, in the data search apparatus according to Item 1, the designation means designates the type of the record and the creating method for creating new data from the record of the type. The generating means stores data created from the record in the first record string stored in the storing means in the record of the type specified by the specifying means according to the creating method specified by the specifying means. Add a field whose field value is
Further, it is characterized in that a new second record string is generated by removing the records of the type not designated by the designation means.

According to a twenty-fourth aspect of the present invention, in the data search apparatus according to the first aspect, the type of the record, the field of the record of the type, and new data from the record of the type are supplied from the designating unit. A creation method to be created is designated, and the generation means sets the field designated by the designation means of the record of the type designated by the designation means of the first record string stored in the storage means,
According to the creating method specified by the specifying means, a new second record string in which the data created from the record is replaced with a field having a field value is generated.

According to the invention described in claim 25, in the data search apparatus according to claim 1, the type of record, the field of the record of the type, and new data from the record of the type are supplied from the designating means. A creation method to be created is designated, and the generation unit specifies the field designated by the designation unit of the record of the type designated by the designation unit of the first record string stored in the storage unit,
A new second record string is created by replacing the data created from the record according to the creating method specified by the specifying unit with a field having a field value, and removing a record of a type not specified by the specifying unit. It is characterized by generating.

[0032]

According to the present invention, since there is a generating means for generating a new record string by processing the record string stored in the storing means based on the processing request inputted from the designating means, Search, concatenation, removal, comparison, orthogonal product, extraction, hierarchical, non-hierarchical, record not only for conventional record type consisting of one kind of record but also for multi-type record string consisting of plural kinds of records It is possible to change the order of the fields inside, add, replace, etc. In addition, by providing some functions of these processes, the user can operate a multi-type record string that represents logical information of a document by specifying a combination of primitive operations that are easy to understand. It is configurable. As a result, like the conventional relational algebra and extended relational algebra, it becomes easy to understand what the designated operation is even in the processing for the multi-type record sequence.

[0033]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT FIG. 1 is a block diagram showing an embodiment of a data search apparatus of the present invention. In the figure, 1 is a storage means, 2 is a designation means, 3 is a generation means, and 4 is a display means. Storage means 1
Stores data to be searched (hereinafter referred to as search target data). The storage unit 1 can be composed of, for example, a hard disk device and an interface. The designating means 2 is for instructing the search content to be requested, and the input search request is input to the generating means 3. The designation means 2 is composed of, for example, an input device such as a keyboard and a mouse, an input control unit, a designation analysis unit that analyzes the input designation. Or, like the storage means 1,
It is also possible to configure by a hard disk device or the like, hold data for instructing search contents, read it out, and input it to the generation means 3. The generation unit 3 reads the search target data designated by the designation unit 2 from the storage unit 1 and executes a search process to generate new search target data. The generating means 3 is composed of a reading unit that reads in search target data, a search processing unit that performs search processing, and the like. The display unit 4 displays the result of the search process performed by the generation unit 3. The display unit 4 includes, for example, a CRT display device, a display control unit, a display information generation unit that generates information displayed on the CRT display from the search target data generated by the generation unit 3. Designating means 2
The designated analysis unit, the display information generation unit of the display unit 4, the reading unit and the search processing unit of the generation unit 3, and the like can be configured by a central control device and a main storage device.

The above-described data search device can handle the logical structure of a document, for example. In this case, the layout of the document can be obtained by applying the format of the document logical structure of the search target data generated by the generation means 3 to the formatter.

The retrieval target data stored in the storage means 1 can be accessed by an identifier. For example, a table including an identifier of the search target data and a pointer to the search target data is stored in the storage unit 1. By reading this table into the main storage device when the data search device is activated, the search target data can be accessed by the identifier after the data search device is activated.

The search target data is composed of a data definition part showing the data structure of the multi-type record string and an actual data part in which the multi-type record string is stored. The data definition part corresponds to a relational database, an extended relational database, or a schema in NST. Also, the actual data part is a relational database, or
It is an expanded relational database or an occurrence in NST.

2 and 3 are explanatory views of the first example of the search target data. 2 shows the data definition part, and FIG. 3 shows the actual data part. In the explanation after FIG. 2,
When expressing the contents of the data definition part and the actual data part, the set is expressed by being enclosed in "{" and "}", and the field name or field value sequence is enclosed by being enclosed in "[" and "]". Has been done. Further, the columns of the record are represented by being surrounded by "<" and ">". The elements of the set or column are separated by commas. The examples shown in FIGS. 2 and 3 represent search target data regarding a certain minutes. This minutes is lined with questions and opinions over time.

One search target data is composed of the data definition part shown in FIG. 2 and the actual data part shown in FIG. The data definition part includes an atomic data definition part, a record definition part, and a route definition part.

The atomic data definition section is a set of atomic data definitions. The atomic data definition defines atomic data that appears in the actual data part, and is composed of a pair of an atomic data type name and an atomic data type. In the following explanation, the atomic data definition is
Atomic data type name, colon, and atomic data type are shown. Atomic data types include string type, integer type, and real number type. For example, the "speaker: character string type" of the atomic data definition part of the example shown in FIG. 2 is one atomic data definition, and the "speaker" of these is the atomic data type name, "character string type". Is an atomic data type. Also, the atomic data type name "contents" is
It is a "string type" atomic data type. In the atomic data definition part, for all atomic data type names appearing in the data definition part, there is an atomic data definition having the atomic data type name. Multiple atomic data definitions with the same atomic data type name must not exist in the atomic data definition part.

The record definition section is a set of record definitions. The record definition defines a record that appears in the actual data section, and is composed of a pair of record type name and record type. In the following description, the record definition is indicated by the record type name, colon, and record type. The record type name is for identifying what kind of record the record definition is about. Also, the record type is composed of a column of field names, and indicates what kind of fields the records of the kind indicated by the record kind name are made of, and in what order the fields are arranged. In the record definition part, for all record type names appearing in the data definition part, there is a record definition having the record type. There cannot be multiple record definitions with the same record type name in the record definition section. In the example shown in FIG. 2, "Question: [speaker, {speak, answer}]" in the record definition section is one record definition, "question" is a record type name, and "[speaker, {speak, Answer}] ”is a record type. In the following description, a record having the record type name “A” will be simply referred to as an A record.

The field name is the atomic data type name,
Alternatively, it is a set of record type names. For example, when the field name is the atomic data name "a",
The value of the field has, as a value, the atomic data type data of the atomic data definition having "a" as the atomic data name. When a field name in a record type is a set of record type names, that field is called a composite field. When the field name is "{b, c}", the value of the field is a multi-type record set consisting of a record having a record kind name of "b" and a record having a record kind name of "c". It is shown that. In the example shown in FIG. 2, the record type name “question” has a value of a character string type field representing “speaker” and a multi-type record sequence including two types of records, “speech” and “answer”. Two fields of a composite field are defined as being arranged in this order. In the following description, a field having the field name “A” is referred to as A
I will call it the field.

The route definition part is a set of record type names. The route definition part indicates what kind of record the actual data part is composed of.
For example, FIG. 2 shows that the actual data part is a set of two types of records, which are a “question” record and an “opinion” record. Note that the atomic data type name and record type name that appear in the data definition section must not overlap.

The actual data part is a set of various kinds of records. A record is composed of a record type name for identifying the type of record and a record body which is a column of field values. In the following description, a record is represented by a concatenation of the record type name, a colon, and the record body. The record body is represented by a field value column enclosed in "[" and "]". For which part of the record body is the field value of which field, which field is the number of fields from the beginning of the record, and what its data type is, the record definition with the same record type name is used. You can know it by referring to it. The order of the field names appearing in the record type of the record definition and the order of appearance of the field values in the record body match. In the example shown in FIG. 3, the actual data part is composed of two question records and three opinion records. The record body of the first question record has two field columns. In this record, the value of the speaker field is "a mountain a'kura", {speak,
The value of the answer field is a multi-type record string consisting of three records.

Next, an example of the processing performed by the generating means 3 will be described. As a first processing example, an example will be described in which a process (selection process) for extracting a record that meets a condition is performed on a multi-type record string. In the first processing example, the generation unit 3 receives the designation of the search target data identifier, the record type name, and the condition for the record of the type from the designation unit 2. These pieces of information are designated by the user via the input device which is a part of the designation means 2.

FIG. 4 is an explanatory diagram of an example of the format of input information when performing the selection process. The outline of the information to be input when the user designates the identifier of the search target data, the record type name, and the condition for the record of the type via the input device of the designation unit 2 is shown. In FIG. 4, the “identifier” is a part that specifies the identifier of the search target data. The “record type name” is a part for specifying the record type name. “Condition” is a part that specifies a condition for a record of the type described before by.

FIG. 5 is an explanatory diagram of an example of input information when performing the selection process. In FIG. 5, it is specified that the search target is the search target data having the identifier “minutes”. Also, the record to be selected is an opinion record, and the selection condition is
The value of the speaker field is "c island c'tree".

Depending on the record, there may be a plurality of fields having the same field name in one record. For example, in the record definition section, "mail:
If there is a record definition "[address, content, address]", the mail record has two address fields. In such a case, when specifying the condition, the field may be specified by "field name.natural number". The natural number part is the field having the specified field name and indicates the order of occurrence. The ".natural number" part may be omitted. When omitted, it is assumed to be the same as "field name. 1". For example, in the “record type name by condition” part of FIG.
By specifying "mail by address.2 = ○○○", the condition that the value of the second address field of the mail record is ○○○ is selected.
It can be specified as the condition of section.

The conditions that can be specified include the following conditions in addition to the conditions shown in FIG.・ Whether the value of the specific field contains a specific character string. · The value of a particular field is greater than a certain number
Is it above, below, or below? • Are the values in the different fields equal? -Is the value (character string) of one field included in the value (character string) of another field? Whether the value (numerical value) of one field is greater than, greater than, less than, or less than, the value (numerical value) of another field. · Whether the value of a composite field contains records that meet certain conditions.・ Are all the elements of the value of a certain composite field satisfying a certain condition? -The logical sum, logical product, negation, and implication of the above conditions.

The designating means 2 receives the input from the user, analyzes it by the designating analysis section, and sends the identifier of the data to be searched, the record type name, and the condition for the record of the type to the generating means 3. After receiving the identifier, the record type name, and the condition from the specifying unit 2, the generating unit 3 reads the search target data having the identifier sent from the specifying unit 2 from the storage unit 1 into the main storage device. Among the elements of the actual data part of the read search target data, a new multi-type in which a record having the record type name sent from the designating means 2 and not satisfying the condition sent from the designating means 2 is removed. Create a record string. Then, the contents of the read data to be searched are defined as the data definition part, and the created new multi-type record string is used as the actual data part to create new search target data. This new search target data is displayed on the CRT display by the display means 4, and the user can see the search result.

FIG. 6 is an explanatory diagram of an example of the actual data part in the new search target data generated by the selection process. In the above selection process, for example, when the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”, the user Performs the input shown in FIG. Then, the data definition part of the created search target data becomes the data definition part shown in FIG. 2 similarly to the data definition part of the original search target data, and the actual data part becomes as shown in FIG. In FIG. 6, from the record shown in FIG. 3, a record that is a designated opinion record and does not meet the condition “speaker = c island c ′ tree”, that is, the speaker is “a mountain a ′ warehouse”. The record is removed and a new actual data part is constructed.

In the above selection process, among the elements of the actual data part of the search target data, the record having the record type name sent from the designating means 2 and not satisfying the condition sent from the designating means 2 A new multi-type record string was created by removing the records, but a record having the record type name sent from the designating means 2 and not satisfying the condition sent from the designating means 2 is sent from the designating means 2. It can also be configured to create a new multi-type record string by removing records that do not have the existing record kind name. FIG. 7 shows s
It is explanatory drawing of another example of the real data part in the new search object data produced | generated by the selection process.
Similar to the case of FIG. 6, for example, when the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user performs the input shown in FIG. In the case of this another example, the actual data part of the created search target data is as shown in FIG. In Figure 7,
From the records shown in FIG. 3, an opinion record that does not meet the condition “speaker = c island c ′ tree” and a record other than the opinion record, that is, an opinion record, "Speaker = c island c '
A new actual data part is composed only of records that conform to "tree".

Next, as a second processing example, an example of performing a processing (union processing) for a multi-type record string and creating a concatenated string in which another multi-type record string is concatenated after the multi-type record string explain.
In the second processing example, the generation unit 3 identifies the identifier of the search target data (referred to as linked search target data) including the actual data part to be connected column from the designating unit 2 and the actual data part to be connected column. It receives the identifier of the search target data (which will be referred to as the linked search target data) including the search target data. These pieces of information are designated by the user via the input device which is a part of the designation means 2. The designation unit 2 receives an input from the user, analyzes the input by the designation analysis unit, and sends the identifiers of the two search target data to the generation unit 3. The generating means 3 is the specifying means 2
After receiving the identifier of the linked search target data and the identifier of the linked search target data from the first, the search target data having the identifier sent from the designating means 2 is first read from the storing means 1 into the main storage device and read. A new data definition part is created from the data definition parts of the two search target data, and it is confirmed whether or not it meets the rules of the data definition part.
If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. If the rule is satisfied, in the multi-type record string of the actual data part of the linked search target data,
Create a new multi-type record string that is a concatenation of the multi-type record strings of the actual data part of the linked search target data. This new multi-type record string is used as the actual data part, and new search target data is created together with the created new data definition part. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

In this second processing example, the union processing can be executed even between search target data having different data definition parts. However, confirm that the new data definition part satisfies the rule, and if it does not, then u
The nion process cannot be executed.

A method of creating a new data definition part will be described.
The atomic data definition part of the new data definition part is created by the union of the atomic data definition parts of the two read target data. Further, the record definition part of the new data definition part is created by the union of the record definition parts of the two read target data. further,
The route definition part of the new data definition part is the union of the route definition parts of the two read target data.

FIGS. 8 and 9 are explanatory views of a second example of the retrieval target data, and FIGS. 10 and 11 are explanatory diagrams of an example of new retrieval target data generated by the union process.
In the union process described above, for example, the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 has the identifier “minutes”, and the data definition part shown in FIG. Then, the search target data having the actual data part shown in FIG. 9 is stored in the storage unit 1 with the identifier "minutes 2".
Suppose it is stored in. At this time, when the user inputs "minutes" as the identifier of the linked search target data and "minutes 2" as the identifier of the linked search target data, the data definition part of the created search target data is shown in FIG.
The data definition part shown in FIG. 11 is obtained, and the actual data part is as shown in FIG.

As a result of the union processing, the new data definition part shown in FIG. 10 is the atomic data definition part of FIG.
By the record definition part and the route definition part and the atomic data definition part, the record definition part and the route definition part of FIG.
The respective union sets are calculated, and the atomic data definition part, the record definition part, and the route definition part of FIG. 10 are generated. Further, in the new actual data part, the record of the actual data part shown in FIG. 9 is concatenated after the record of the actual data part shown in FIG. 3, and the actual data part shown in FIG. 11 is generated. The new data definition section shown in FIG. 10 and the new actual data section shown in FIG. 11 constitute new search target data.

As a third processing example, a multi-type record string is targeted, and from the elements of the multi-type record string,
An example of performing a process (intersection process) of removing a non-element of another multi-type record string will be described. Here, the former multi-type record sequence is referred to as a column to be removed, and the latter multi-type record sequence is referred to as a removal column.

In the third processing example, the generating means 3 wants to specify from the designating means 2 the identifier of the search target data (referred to as the search target data to be removed) including the actual data portion to be the column to be removed and the removal column. An identifier of search target data (referred to as removal search target data) including the actual data part is received.
These pieces of information are designated by the user via the input device which is a part of the designation means 2. The generation unit 3 receives the identifier of the data to be removed to be searched and the identifier of the removal search target data from the designation unit 2 and then stores the search target data having the identifier sent from the designation unit 2 to the storage unit 1
Read from main memory. A new data definition part is created from the data definition parts of the two read target data.
As a result, if a new data definition part cannot be created, or if the created new data definition part does not satisfy the rules of the data definition part, a display command of a data definition non-conformance message is sent to the display means 4, and the processing is performed. finish.
Further, a new multi-type record string is created by removing the records that are not included in the multi-type record string of the real data part of the data to be removed and searched from the read multi-type record string of the real data part of the data to be removed.
The created new multi-type record string is used as a new actual data part, and new search target data is configured together with a new data definition part. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

In the third processing example, the intersection process can be executed even between the search target data having different data definition parts. However, it is confirmed that the new data definition part satisfies the rule, and if the rule is not satisfied, the intersection process cannot be executed.

How to create a new data definition part will be described.
First, it is confirmed that there are no two atomic data definitions having the same atomic data type name but different atomic data types in the data definition parts of the two retrieved search target data. If it exists, a new data definition cannot be created. Next, it is confirmed that two record definitions having the same record type name but different record types do not exist in the data definition parts of the two retrieved search target data. Also in this case, if it exists, a new data definition part cannot be created. Furthermore, it is confirmed that the same name does not appear in the atomic data type name and the record type name in the data definition parts of the two retrieved search target data. If it exists, a new data definition cannot be created. The atomic data definition part is created by the intersection of the atomic data definition parts of the data definition parts of the two retrieved search target data. The record definition part is created by a product set of the record definition parts of the data definition parts of the two read target data. The route definition part is created by a product set of the data definition parts of the two retrieved search target data.

FIG. 12 is an explanatory diagram of a third example of the actual data part of the retrieval target data, and FIGS.
It is explanatory drawing of an example of the new search object data produced | generated by the action process. The intersectio mentioned above
In the n process, for example, the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 has the identifier “minutes”, and the data definition part shown in FIG. Consider a case where the search target data having the actual data part shown in FIG. 12 is stored in the storage means 1 with the identifier "minutes 2". At this time, when the user inputs “minutes” as the identifier of the data to be removed to be searched and “minutes 2” as the identifier of the data to be removed to be searched, the search target data to be created is the data definition section shown in FIG. Then, new search target data composed of the actual data part shown in FIG. 14 is generated. In the atomic data definition part shown in FIG. 2 and FIG. 8, there is no atomic data definition with the same atomic data type name but different atomic data types, and there is no record definition with the same record type name but different record types. It does not exist, and the same name does not appear in the atomic data species name and the record species name. From these confirmations, the data definition part shown in FIG. 2 and FIG.
Rsection processing is possible. A product set is created for each of the atomic data definition part, the record definition part, and the route definition part, and a new data definition part shown in FIG. 13 is created. In the actual data part, the records not included in the multi-type record string of the actual data part shown in FIG. 12 are removed from the multi-type record string of the actual data part shown in FIG. A department is created.
New search target data is configured by the data definition section shown in FIG. 13 and the actual data section shown in FIG.

As a fourth processing example, a multi-type record string (referred to as a compared column) targeted for a multi-type record string is changed to another multi-type record string (referred to as a comparison column). Processing to remove elements (dif
An example of performing a "fence process" will be described. In the fourth processing example, the generation unit 3 receives from the designation unit 2
The identifier of the search target data (called the compared search target data) that includes the actual data part that you want to make the compared column and the identifier of the search target data (called the comparative search target data) that contains the actual data part that you want to make the comparison column receive. These pieces of information are designated by the user via the input device which is a part of the designation means 2. After receiving the identifier of the compared search target data and the identifier of the comparative search target data from the specifying unit 2, the generating unit 3 stores the search target data having the identifier sent from the specifying unit 2 from the storage unit 1 in the main memory. Read into the device. For the data definition sections of the two retrieved search target data,
Confirm whether the following conditions (1) to (3) are satisfied. If the result of the confirmation is that the conditions are not satisfied, a command to display a data definition inconsistency message is sent to the display means 4, and the process ends. (1) Two atomic data definitions having the same atomic data type name but different atomic data types do not exist in the data definition parts of the two retrieved search target data. (2) Two record definitions having the same record type name but different record types do not exist in the data definition parts of the two retrieved search target data. (3) The same name does not appear in the atomic data type name and the record type name in the data definition parts of the two retrieved search target data. When it is confirmed that these conditions are satisfied, the new type is obtained by removing the records contained in the multi-type record string of the actual data part of the comparative search target data from the multi-type record string of the actual data part of the compared search target data. Create a multi-type record column. New search target data is constructed in which the created new multi-type record string is used as a new actual data part and the data definition part of the compared search target data is used as a new data definition part. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

In the fourth processing example, the difference processing can be executed even between search target data having different data definition parts. However, if it is confirmed that the above conditions (1) to (3) are satisfied and the new data definition part does not satisfy the rule, the difference process cannot be executed.

FIG. 15 is an explanatory diagram of an example of the actual data part of the new search target data generated by the difference process. In the difference processing described above, for example, the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 has the identifier “minutes”, and the data definition part shown in FIG. And Fig. 1
Consider a case where the search target data having the actual data part shown in 2 is stored in the storage means 1 with the identifier "minutes 2". At this time, when the user inputs "minutes" as the identifier of the compared set search target data and "minutes 2" as the identifier of the comparison set search target data, the data definition unit shown in FIG. New search target data composed of the actual data portion shown is generated. 2 and 8
In the atomic data definition part shown in, there is no atomic data definition with the same atomic data type name but different atomic data types, and there is no record definition with the same record type name but different record types. , The same name does not appear in atomic data type name and record type name. From these confirmations, the difference processing can be performed on the two multi-type record strings. The record of the actual data part shown in FIG. 12 is removed from the record of the actual data part shown in FIG. 3 to create a new actual data part shown in FIG. By the data definition part of the compared search target data shown in FIG. 2 and the actual data part shown in FIG.
New search target data is constructed.

As a fifth processing example, an element of a certain multi-type record string (this is called a direct product column) for another multi-type record string is another multi-type record string (this is called a direct product column). ) Elements can be connected,
Process to create a column consisting of all records (cartesi
An example of performing an product process) will be described. Here, the concatenation of two records is to create a new record including the fields of all the two records. The target of the cartesian product process is a multi-type record string having m elements and n
If it is a multi-type record string having elements, the result of the cartesian product process,
A multi-type record string having m × n elements is generated.

In the fifth processing example, the generating means 3 wants the identifier of search target data (called direct product search target data) including the actual data portion desired to be the direct product sequence and the direct product sequence from the designating means 2. An identifier of search target data (referred to as direct product search target data) including an actual data part is received.
These pieces of information are designated by the user via the input device which is a part of the designation means 2. After receiving the identifier of the direct product search target data and the identifier of the direct product search target data from the designating means 2, the generating means 3 stores the search target data having the identifier sent from the designating means 2 from the storage means 1 in the main memory. Read into the device. A new data definition part is created from the data definition parts of the two read target data, and it is confirmed whether or not it meets the rules of the data definition part. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends.

Next, from all combinations of the records that are the elements of the real data part of the data to be searched for and the records that are the elements of the real data part of the product to be searched,
Create a new multitype record column. The new multi-type record string created is used as the actual data part, and new search target data is created along with the new data definition part. The new search target data is displayed on the display unit 4 as C.
Displayed on the RT display, the user can see the search results.

How to create a new data definition part will be described.
The atomic data definition part is created by the union of the atomic data definition parts of the data definition parts of the two read target data.

The record definition part is a set of record definition parts for the data to be searched for, a set of record definition parts for the data to be searched for direct product, and a record definition for the data to be searched for direct product. For each combination of the element that is the element of the route definition part of the search target data and the record definition of the direct product search target data whose record type name is the element of the route definition part of the direct product search target data On the other hand, it is the union of three sets of record definition created by performing the following processes (1) to (3). (1) A new record type name created by concatenating "." And the record type name of the record definition of the data to be searched for is added after the record type name of the record definition of the product to be searched for to be combined. create. (2) A new record type is created by concatenating the record types of the record definition of the product to be searched to be combined after the record type of the record definition of the data to be searched for to be combined. (3) Create a new record definition with the result of (1) as the record type name and the result of (2) as the record type.

The route definition part has a "." After the record type name which is an element of the route definition part of the data to be searched.
And a record type name that is an element of the route definition part of the direct product search target data.

Next, a method of creating a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following (1) to (3) processes are performed on the amd product column of the real data part of the amd product search target data to create a set of records until the processes are performed. (1) The first record (called a direct product record) of the direct product column is extracted. In the following description, when it is stated that a record is extracted from the multi-type record string, the extracted record is assumed to be deleted from the multi-type record string. (2) Prepare a temporary record string 2 which is a multi-type record string having no elements. For each record from the beginning to the end of the direct product sequence of the real data part of the direct product search target data (this is called a direct product record), “.” Is added in order after the record type name of the direct product record. Create the record kind name by concatenating the record kind name of the Cartesian product record, create the record body by connecting the record body of the Cartesian product record after the record body of the Cartesian product record, and the created record kind name, record body The record constituted by is connected to the end of the temporary record string 2. By this processing, the tentative record string 2 becomes a record string created by all combinations of a certain direct product record and the direct product record of the direct product search target data. (3) The record string resulting from (2) is linked to the end of the temporary record string 1. When the process is performed on all the elements of the direct product sequence, a new multi-type record sequence is created in the temporary record sequence 1. This new multi-type record string becomes the new actual data part.

16 and 17 show the fourth search target data.
Of the example of FIG. 18 and FIG.
It is explanatory drawing of an example of the new search object data produced | generated by the product process. Above cartesia
In the n product process, for example, the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 has the identifier “minutes”, and FIG.
Consider a case where the search target data having the data definition part shown in FIG. 4 and the actual data part shown in FIG. 17 is stored in the storage means 1 by the identifier “participant list”. At this time,
When the user inputs “minutes” as the identifier of the direct product search target data and “participant list” as the identifier of the direct product search target data, the search target data created is the data definition unit shown in FIG. New search target data composed of the actual data part shown in FIG. 19 is generated. Figure 2
The atomic data definition part shown in FIG. 16 and the new atomic data definition part shown in FIG. 18 are created by the union. The new record definition part is shown in FIGS.
It is composed of the record definition of the record definition part shown in and the record definition created by the combination thereof. For example, in FIG. 18, a combination of the record definition “question: [speaker, {speak, answer}] in FIG. 2 and the record definition“ in-house: [name, affiliation] in FIG. -A new record definition with "in-house" and record type "[speaker, {speak, answer}, name, department]" is created. A new route definition part is also created by combining the elements of the route definition part of FIG. 2 and the elements of the route definition part of FIG. These new atomic data definition part, record definition part, and route definition part constitute a new data definition part. Further, the record string of the actual data part shown in FIG. 3 and the record string of the actual data part shown in FIG. 17 are also combined in order,
A new record is created, and a new actual data part as shown in FIG. 19 is created by the sequence of the new record. The new search target data is composed of the created new data definition part and the new actual data part.

As a sixth processing example, an example will be described in which processing for changing the arrangement order of fields of a record is performed for a multi-type record string. In the sixth processing example, the generation unit 3 receives from the designation unit 2 an identifier of search target data, a record type name, a method of rearranging fields of records of the type, and a designation of a record type name after rearrangement. These pieces of information are designated by the user via the input device which is a part of the designation means 2.

FIG. 20 is an explanatory diagram showing an example of the format of input information when performing the rearrangement process. User specifies 2
Via the input device of, search target data identifier, record type name, how to sort the fields of the record of the type,
It shows an outline of the information to be input when specifying the record type name after sorting. In FIG. 20, "identifier"
Is a part for designating the identifier of the search target data.
The “record type name” is a part for specifying the record type name. The “record type name after sorting” is a part that specifies the record type name of the record after sorting. In "field designation", a field is designated. The method of specifying the field is the same as the method of specifying the condition in the first processing example. "Field specification"
Indicates how to sort the fields of the record of the type specified by the "record type name". The order of "field designation" indicates the order of fields after rearrangement.

FIG. 21 is an explanatory diagram of an example of input information when performing the rearrangement process. In FIG. 21, it is specified that the search target is the search target data having the identifier “minutes”. The record to be rearranged is a question record, and the field order of the question record is specified to be rearranged in the order of {speak, answer} field and speaker field. It is also specified that the new record type name after rearrangement is “Question 2”.

The generating means 3 has the identifier sent from the designating means 2 after receiving from the designating means 2 the identifier of the data to be searched, the record type name and the designation of the rearrangement of the fields of the records of the type. The search target data is read from the storage means 1 into the main storage device. Create a new data definition part from the data definition part of the read search target data,
Make sure it meets the rules of the Data Definition Division. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. After confirming that the rules of the data definition section are satisfied, rearrange the field order of the record body of the specified type record among the elements of the actual data section of the read target data to the specified field order. , Furthermore, a new multi-type record string is created in which the record type name is changed to the specified sorted record type name. The created new multi-type record string is used as a new actual data part, and new search target data is created along with a new data definition part. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

A method of creating a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the record definition part of the retrieved search target data.
A record definition having a record type name specified by "record type name after rearrangement" and a field type of a field specified by "field specification" arranged in the specified order is added. The route definition part removes the one specified by "record type name" from the route definition part of the read target data and adds the record type name specified by "record type name after sorting". .

22 and 23 are explanatory views of an example of new search target data generated by the rearrangement process.
In the above-mentioned rearrangement process, for example, consider a case where search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. . At this time, when the user inputs the rearrangement instruction shown in FIG. 21, the search target data created is a new search target composed of the data definition part shown in FIG. 22 and the actual data part shown in FIG. Data is generated. As shown in FIG. 22, in the new data definition section, the definition of the question 2 record is added. Also, FIG.
As shown in, the question record in the new actual data part is the question 2 record. The order of the fields of the question 2 record is {statement, answer}, as specified.
The fields are rearranged in the order of the speaker field.

In the above-mentioned rearrangement processing, among the elements of the actual data part of the retrieval target data, the field order of the record body of the record of the specified type is rearranged in the specified field order, and the record type name is specified. The new real data part was constructed by the new multi-type record string that was changed to the sorted record type name, but the new real data part was constructed by the new record string created only from the specified type of record. It is also possible. That is, among the elements of the actual data part of the search target data, the field order of the record body of the record of the specified type is rearranged in the specified field order, and the record type name is the specified record type name after the rearrangement. Change to. Further, it can be configured to remove records other than the designated type to create a new multi-type record sequence. FIG. 24 is an explanatory diagram of another example of the actual data part in the new search target data generated by the rearrangement processing. Similar to the case of FIG. 23, for example,
When the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”, the user inputs the input shown in FIG. Consider when to do it. In this case, the actual data part of the created search target data is as shown in FIG. In FIG. 24, the order of the fields of the question record is changed from the record shown in FIG. 3, and the multi-type record string is created only by the question 2 record with the record type name “question 2” after rearrangement. A real data part.

As a seventh processing example, an example of performing a process (projection process) for extracting a field from a record for a multi-type record string will be described. In the seventh processing example, the generation unit 3 receives the identifier of the search target data, the record type name,
The designation of the field to be extracted from the record of the type and the record type name after the extraction is received. These pieces of information are designated by the user via the input device which is a part of the designation means 2.

FIG. 25 is an explanatory diagram of an example of the format of input information when performing the extraction process. An outline of information to be input when the user designates the identifier of the search target data, the record type name, the field to be extracted from the record of the type, and the record type name after the extraction through the input device of the designation means 2 is shown. ing. In FIG. 25, the “identifier” is a part that specifies the identifier of the search target data. The “record type name” is a part for specifying the record type name. The “record type name after extraction” is a part that specifies the record type name of the record after the field is extracted. In "field designation", a field is designated. The method of specifying the field is the same as the method of specifying the condition in the first processing example. The "field designation" is for indicating the field extracted from the record of the type designated by the "record type name". The order of "field specification" indicates the order of the fields after extraction.

FIG. 26 is an explanatory diagram of an example of input information when performing the extraction process. In FIG. 26, it is specified that the search target is the search target data having the identifier “minutes”. Further, it is specified that the record from which the field is to be extracted is the question record, and the field extracted from the question record is the {speech, answer} field. It is also specified that the new record type name after extraction of the field is “Question 2”.

The generation means 3 receives from the designation means 2 the identifier of the data to be searched, the record type name, the designation of the field to be extracted from the record of the type, the record type name after extraction, and then sends from the designation means 2. The retrieval target data having the obtained identifier is read from the storage means 1 into the main storage device. A new data definition part is created from the data definition part of the read search target data, and it is confirmed whether or not it meets the rules of the data definition part. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. After confirming that the rules of the data definition section are satisfied, a new actual data section is created from the actual data section of the retrieved search target data. In the created new actual data part, new search target data is created together with the new data definition part. The new search target data is displayed on the display unit 4 as C.
Displayed on the RT display, the user can see the search results.

A method of creating a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the record definition part of the retrieved search target data.
A record definition having a record type name specified by "record type name after extraction" and a field type of the field specified by "field specification" arranged in the specified order is added. The route definition part is a route definition part of the read search target data, in which the one designated by "record species name" is removed and the record species name designated by "record species name after extraction" is added.

A method of creating a new actual data part will be described. Among the elements of the actual data part of the read search target data, a new record body created by extracting the specified field from the record body of the record of the specified type, and the "record type name after extraction" Create a new multi-type record string that is replaced with a new record consisting of the record type name specified in. At this time, the order of the fields of the new record body follows the order in which the fields are specified. A new actual data part is configured by the created new multi-type record string.

27 and 28 are explanatory views of an example of new search target data generated by the extraction processing. In the above-mentioned extraction processing, for example, consider a case where the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. . At this time, when the user inputs the extraction instruction shown in FIG. 26, the search target data created is
New search target data composed of the data definition part shown in FIG. 27 and the actual data part shown in FIG. 28 is generated. As shown in FIG. 27, in the new data definition part, the definition of the question 2 record is added and the route definition part is changed. Also, as shown in FIG. 28, in the new actual data part, the {statement, answer} field is extracted from the question record, the body of the record is replaced, and the record type name is replaced with "question 2". .

In the above extraction process, the record of the specified type among the elements of the actual data part of the search target data is
New multi-type record string replaced with a new record consisting of a new record body created by extracting the specified field from the record body of the record and the record kind name specified in "Record kind name after extraction" However, it is possible to configure to create a new multi-type record string created only from the specified type of record. That is, a new record body created by extracting the designated field from the record body of the record of the designated type,
It can also be configured to create a new multi-type record column by replacing with a new record consisting of the record type name specified in "Record type name after extraction" and removing the records of the type not specified. it can. FIG. 29 is an explanatory diagram of another example of the actual data part in the new search target data generated by the extraction processing. As in the case of FIG. 28, for example, when the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user performs the input shown in FIG. In this case, the actual data part of the created search target data is as shown in FIG. In FIG. 29,
Of the records shown in FIG. 3, the specified field is extracted from the question record, and only the question 2 record with the extracted record type name “question 2” constitutes the new actual data part. .

As an eighth processing example, a first hierarchical processing of records for a multi-type record string is performed.
An example will be described. The record hierarchization process is a process of collecting records having the same value of a specific field (hereinafter referred to as a key field) into one new record. Multiple fields may be key fields. The new record has a key field and a composite field whose value is a multi-type record string whose elements are records having the same key field value.

In the eighth processing example, the generation means 3 receives from the designation means 2 the identifier of the search target data, the record type name, and the designation of the field to be the key field. These pieces of information are designated by the user via the input device which is a part of the designation means 2.

FIG. 30 is an explanatory diagram of an example of the format of input information when performing the hierarchical processing. The outline of the information to be input when the user designates the identifier of the search target data, the record type name, and the field name to be the key field via the input device of the designating means 2 is shown. In FIG. 30, the "identifier" is a part that specifies the identifier of the search target data. The “record type name” is a part for specifying the record type name. In "field designation", a field is designated. The method of specifying the field is the same as the method of specifying the condition in the first processing example. The "field designation" indicates a key field among the fields of the record of the type designated by the "record type name". However, the record of the type specified by "record type name" must include the field specified as the key field.

FIG. 31 is an explanatory diagram of an example of input information when performing the hierarchical processing. In FIG. 31, it is specified that the search target is the search target data having the identifier “minutes”. The records to be layered are question records and opinion records.
The field is specified to be the speaker field.

The generation means 3 receives the identifier of the retrieval target data, the record type name, and the designation of the field to be the key field from the designation means 2, and then the retrieval target data having the identifier sent from the designation means 2. Is read from the storage means 1 into the main storage device. A new data definition part is created from the data definition part of the read search target data, and it is confirmed whether it satisfies the rules of the data definition part. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. After confirming that the rules of the data definition section are satisfied, a new actual data section is created from the actual data section of the retrieved search target data. In the created new actual data part, new search target data is created together with the new data definition part. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

How to create a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the record definition part of the retrieved search target data.
The record type name is a concatenation of the record type names of the type designated by the designating means 2 in the designated order by ".", And the record type is the key field of the record of the type designated by the designating means 2. Field name and specifying means 2
A record definition consisting of fields of a set of record type names sent from is added. The order of the field names is the order of the set of the field name of the key field and the specified record type name. The route definition part is obtained by removing the specified "record type name" from the route definition part of the read search target data and adding the specified record type names concatenated with "." In the specified order. Is.

A method of creating a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following processes (1) and (2) are executed until there are no elements in the actual data part of the search target data. (1) The first record is extracted from the actual data part of the search target data. (2) If the record fetched in (1) is of a type not designated by the designating means, the record is linked to the end of the temporary record sequence 1 and the process returns to (1). If not, a temporary record string 2 which is a multi-type record string having no element is prepared and the record is linked to the end of the temporary record string 2. After that, the records resulting from the following processes (a) and (b) are connected to the end of the temporary record string 1, and the process returns to (1). I. The first record of the actual data part of the search target data is
It is a record of the type specified by the specifying means, and
The value of the key field of the record is the temporary record string 2
When the value is equal to the value of the key field of the last record of, the first record is extracted from the actual data part of the search target data, linked to the last of the temporary record sequence 2, and the process of b is repeated. In the other cases, the process of B is executed. B. The record type name specified by the specifying means 2 is connected by "." To form a record type name, which is composed of the key field of the first record of the temporary record string 2 and the field whose value is the temporary record string 2. Create a new record consisting of a new record body. The order of concatenation of record kind names follows the specified order. The order of the fields of the new record body is the order of the key field and the non-key field.

The process (a) is a process of using consecutive records having the same key field value as a multi-type record string, and constitutes a non-key field of the new record body to be created. In addition, the process of (b) is a process of making the multi-type record string created in the temporary record string 2 so far into one record when the values of the key fields are not continuous. Will be linked to the temporary record sequence 1. The temporary record string 1 becomes a new actual data part by executing the above-described processing until the elements of the actual data part of the search target data are exhausted.

32 and 33 show the fifth search target data.
34 and 35 are explanatory diagrams of an example of new search target data generated by the first hierarchical processing. In the hierarchical processing described above, for example, the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG.
Suppose it is stored in. At this time, the user
When the hierarchization instruction shown in FIG. 31 is input, the search target data created are the data definition section shown in FIG.
New search target data composed of the actual data part shown in FIG. 5 is generated. As shown in FIG. 34, the new data definition section has a record type name of “question.
A record definition of a record type including a set of record type names designated as a key field "speaker" is added. In addition, the new route definition part deletes the specified record type names "question" and "opinion", and connects the specified record type names with "." In the specified order to form the "question.opinion" element. Has been added. As shown in FIG. 35, the actual data section collects consecutive records having the same speaker field from the designated question record and opinion record, and sets the speaker field and the row of the collected records as a composite field. Question. Creating an opinion record.

In the above-described hierarchical processing, a new actual data part is constructed by a record of a type not specified among the elements of the actual data part of the retrieval target data and a new record created from the record of the specified type. However, it is also possible to form a new actual data part only by a new record created from a specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is of a type not specified by the specifying means, the process returns to (1). Constitute. This will create only a new record from the actual data part of the retrieved search data, consisting of a composite field whose value is the column of consecutive records where the value of the key field is the same as the key value. The actual data part can be constructed by a new multi-type record string having.

FIG. 36 is an explanatory diagram of another example of the actual data part in the new search target data generated by the first hierarchical processing. As in the case of FIG. 35, for example, when the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user performs the input shown in FIG. In this case, the actual data portion of the created search target data is as shown in FIG. In FIG. 36, the question. A new actual data part is composed of a multi-type record string consisting only of opinion records.

In the above two hierarchized examples, only fields common to several types of hierarchized records can be designated as key fields. However, it can be configured so that a different key field can be specified for each record type. However, at this time,
The key field specified for each record type must have the same atomic data type.

As a ninth processing example, a second example of the record hierarchization processing will be described. In this ninth processing example, almost the same processing as in the eighth processing example is performed, but the method of creating the data definition part and the actual data part is different. The difference will be described below.

How to create a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the union of the following three sets. (1) A set of record definitions of the read search target data. (2) Among the record definitions of the retrieved search target data, for each record definition for which the record type name is designated by the designating means 2, a new record type name obtained by concatenating “−” to the record type name is provided, In addition, it is a set of record definitions having a new record type created by removing the field name that is the key field from the record type of the record definition. (3) A new record type name obtained by concatenating the record type names designated by the designating unit 2 in the designated order with ".", And the field name of the key field and the record designated by the designating unit 2 It is a set that has as its element a single record definition that has a new record type in which the seed name is concatenated with "-" arranged in the specified order. The order of the field names is the field name of the key field and the field name of the non-key field. The route definition part removes the specified record type name from the route definition part of the read target data and adds an element that connects the record type names of the specified type with "." In the specified order. Is.

Next, how to create a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following (1) and (2) are applied until there are no elements in the actual data part of the search target data.
The process of is executed. (1) The first record is extracted from the actual data part of the search target data. (2) If the record fetched in (1) is of a type not designated by the designating means, the record is linked to the end of the temporary record sequence 1 and the process returns to (1). If not, a temporary record string 2 which is a multi-type record string having no element is prepared, and the record and the temporary record string 2 are connected to the end of the temporary record string 2. After that, the records resulting from the following processings A to C are connected to the end of the temporary record string 1, and the process returns to (1). I. The first record of the actual data part of the search target data is
When the type of record is designated by the designating means 2 and the value of the key field of the record is equal to the value of the key field of the last record of the temporary record sequence 2, the actual data of the search target data The first record is taken out from the copy, linked to the end of the temporary record string 2, and the process of b is repeated. In the other cases, the process of B is executed. B. A multi-type record in which the record body of each record in the provisional record string 2 is replaced with the record body obtained by removing the key field from the record body, and the record type name is changed to a concatenation of the record type name and "-". Make a line. After that, the processing of c is executed. C. The record type name specified by the specifying means is concatenated with "." To be the record type name. Key field of the first record of the tentative record sequence 2 before the execution of the above process, and the process b. Create a new record with a new record body consisting of fields whose values are the new multi-type record string generated in. The order of concatenation of record kind names follows the specified order. The order of the fields is the order of the key field and the non-key field.

This a. The process of (1) is a process of setting consecutive records having the same key field value as a multi-type record string. The field that is not the key field of the new record body to be created is processed by the above process. Also,
B. , C. Is a process for making the multi-type record string created in the temporary record string 2 so far into one record when the values of the key fields are not continuous, and one created record is a temporary record. It will be linked to the record string 1. It differs from the eighth processing example in that the key field is removed from each record of the temporary record sequence 2. By executing the above processing until there are no elements in the actual data part of the search target data,
The temporary record string 1 becomes the new actual data part. New search target data is created from the new data definition part generated as described above and the new actual data part.

37 and 38 are explanatory views of an example of new search target data generated by the second hierarchical processing. In the second layering process described above, for example, as shown in FIG.
Consider a case where the search target data having the data definition part shown in FIG. 3 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes”. At this time, when the user inputs the hierarchization instruction shown in FIG. 31, the search target data created is the data definition section shown in FIG.
New search target data composed of the actual data part shown in FIG. 38 is generated. As shown in FIG. 37, in the record definition section of the new data definition section, based on the specified record type names "question" and "opinion", question-record, opinion-
Records, questions. Opinion records have been added. The question-record and opinion-record are the original question record,
It has a record type in which the key field “speaker” is deleted from the record type of the opinion record. Also, the question. The opinion record contains the key field "Speaker",
It has a record type consisting of a set of new record type names "question-" and "opinion-" in which "-" is added to the designated record type name. Also, the new route definition part
The specified record type names "question" and "opinion" are deleted, and the specified record type names are "." In the specified order.
The element of "Question. Opinion" linked by is added. As shown in FIG. 38, the actual data section extracts a continuous record having the same speaker field, for example, "a mountain a'kura" from the specified question record and opinion record, and the speaker field "a mountain""a'zo", the record in which the speaker field is deleted from the extracted record and [-] is added to the record type name, for example, "question-: [...]" or "opinion-: [...]". A record having a record body composed of a composite field composed of a multi-type record string of "" is created.

Also in the second example of the layering process described above,
It is also possible to construct a new actual data part only by a new record created from the specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is a record of a type not specified by the specifying means, the process returns to (1). To configure. As a result, from the actual data part of the read search target data,
A new multitype that extracts a column of consecutive records where the value of the key field is the same as the key value, creates a record with the key field and the fields of the multitype record column of the extracted record, and has only this new record The record string can form the actual data part.

FIG. 39 is an explanatory diagram of another example of the actual data part in the new search target data generated by the second hierarchical processing. As in the case of FIG. 35, for example, when the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user performs the input shown in FIG. In this case, the actual data part of the created search target data is as shown in FIG. In FIG. 39, among the records shown in FIG. 38, question. A new actual data part is composed of only the opinion record.

Furthermore, also in this second hierarchical example, the fields that can be designated as key fields are:
Only fields that are common to several types of records that are hierarchical. However, it can be configured so that a different key field can be specified for each record type. However, at this time, the atomic data type of the key field specified for each record type must be the same.

As the tenth processing example, the third example of the record hierarchical processing will be described. In the tenth processing example, almost the same processing as the eighth processing example is performed, but the method of creating the data definition part and the actual data part is different. The difference will be described below.

How to create a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition section is obtained by concatenating the record definition of the read search target data and the record type name of the type whose record type name is designated by the designating means 2 by ".", And the record type is a key field field. It consists of a record definition that is a sequence of names and a set of specified record species names. Here, the order of concatenation of the record type names follows the specified order. The order of the field names is the order of the field name of the key field and the set of the designated record type names. The route definition part is obtained by removing the specified record type name from the route definition part of the read target data and adding an element in which the specified record type name is connected by "." In the specified order. .

Next, how to create a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following (1) and (2) are applied until there are no elements in the actual data part of the search target data.
The process of is executed. (1) The first record is extracted from the actual data part of the search target data. (2) If the record fetched in (1) is of a type not designated by the designating means, the record is linked to the end of the temporary record sequence 1 and the process returns to (1). If not, a temporary record string 2 which is a multi-type record string having no element is prepared and the record is linked to the end of the temporary record string 2. After that, the following a. , B. The record obtained as a result of performing the process of (1) is linked to the end of the temporary record string 1, and the process returns to (1). I. Of the records in the actual data part of the search target data, the record is of the type designated by the designating means 2, and
The value of the key field of the record is the temporary record string 2
When it is equal to the value of the key field of the last record of the record, that record is taken out from the actual data part of the search target data, and is linked to the last of the temporary record sequence 2. Otherwise, do nothing. This process is sequentially executed from the beginning to the end of the actual data part of the search target data. B. The record type name specified by the specifying means 2 is connected by "." To form a record type name, which is composed of the key field of the first record of the temporary record string 2 and the field whose value is the temporary record string 2. Create a new record from the new record body. The order of concatenation of record kind names follows the specified order. The order of the fields is the order of the key field and the non-key field.

The process (a) is a process for extracting the records having the same key field value in the order of the record strings to form a multi-type record string, and constructing a field that is not a key field of the new record body to be created. To do. In addition, the process of (b) is a process for making the multi-type record string created in the temporary record string 2 into one record, and one created record is linked to the temporary record string 1. It differs from the eighth processing example in that records having the same key field are extracted even if they are not continuous. The temporary record string 1 becomes a new actual data part by executing the above-described processing until the elements of the actual data part of the search target data are exhausted.
New search target data is created from the new data definition part generated as described above and the new actual data part.

FIG. 40 is an explanatory diagram of an example of the actual data part in the new search target data generated by the third hierarchical processing. In the above-described third hierarchical processing, for example, the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes”. Consider the case. At this time, when the user inputs the hierarchization instruction shown in FIG. 31, the search target data to be created is a new search composed of the data definition part shown in FIG. 34 and the actual data part shown in FIG. Target data is generated. As shown in FIG. 40,
From the designated question record and opinion record, all the records having the same speaker field, for example, "a mountain a'kura" are sequentially extracted, and the speaker field "a mountain a '
A record having a record body composed of “Kura” and a multi-type record string composed of the extracted records is created.

Also in the third example of the layering process described above,
It is also possible to construct a new actual data part only by a new record created from the specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is of a type not specified by the specifying means, the process returns to (1). Constitute. As a result, all records in which the value of the key field is the same as the key value are extracted from the actual data part of the retrieved search target data, and the key field and the multi-type record string of the extracted record are used as fields. It is possible to create a new record and configure the actual data part with a new multi-type record string having only this new record.

FIG. 41 is an explanatory diagram of another example of the actual data part in the new search target data generated by the third hierarchical processing. As in the case of FIG. 40, for example, when the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user performs the input shown in FIG. In this case, the actual data part of the created search target data is as shown in FIG. In FIG. 41, the question. A new actual data part is composed of only the opinion record.

Further, also in this third hierarchical example, the fields that can be designated as key fields are:
Only fields that are common to several types of records that are hierarchical. However, it can be configured so that a different key field can be specified for each record type. However, at this time, the atomic data type of the key field specified for each record type must be the same.

As the eleventh processing example, the fourth example of the record hierarchical processing will be described. In the eleventh processing example, almost the same processing as in the eighth processing example is performed, but the method of creating the data definition part and the actual data part is different. The difference will be described below.

A method of creating a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the union of the following three sets. (1) A set of record definitions of the read search target data. (2) Among the record definitions of the retrieved search target data, for each record definition for which the record type name is designated by the designating means 2, a new record type name obtained by concatenating “−” to the record type name is provided, In addition, it is a set of record definitions having a new record type created by removing the field name that is the key field from the record type of the record definition. (3) A new record type name obtained by concatenating the record type names designated by the designating unit 2 in the designated order with ".", And the field name of the key field and the record designated by the designating unit 2 It is a set that has as its element a single record definition that has a new record type in which the seed name is concatenated with "-" arranged in the specified order. The order of the field names is the field name of the key field and the field name of the non-key field. The route definition part removes the specified record type name from the route definition part of the read target data and adds an element that connects the record type names of the specified type with "." In the specified order. Is.

Next, a method of creating a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following (1) and (2) are applied until there are no elements in the actual data part of the search target data.
The process of is executed. (1) The first record is extracted from the actual data part of the search target data. (2) If the record fetched in (1) is of a type not designated by the designating means, the record is linked to the end of the temporary record sequence 1 and the process returns to (1). If not, a temporary record string 2 which is a multi-type record string having no element is prepared and the record is linked to the end of the temporary record string 2. After that, the records resulting from the following processings A to C are connected to the end of the temporary record string 1, and the process returns to (1). I. The record of the actual data part of the search target data is a record of the type designated by the designating means 2, and the value of the key field of the record is the key field of the last record of the temporary record sequence 2. When equal to the value,
The record is taken out from the actual data part of the search target data, linked to the end of the temporary record string 2, and the process of b is repeated. Otherwise, do nothing. This process is sequentially executed from the beginning to the end of the actual data part of the search target data. B. A multi-type record in which the record body of each record in the provisional record string 2 is replaced with the record body obtained by removing the key field from the record body, and the record type name is changed to a concatenation of the record type name and "-". Make a line. After that, the processing of c is executed. C. A record type name obtained by concatenating the record type names designated by the designating means with "." Is used as the record type name, and the key field of the first record of the temporary record sequence 2 before the process b and the process b. Create a new record with a new record body consisting of fields whose values are the new multi-type record string generated in. The order of concatenation of record kind names follows the specified order. The order of the fields is the order of the key field and the non-key field.

The process (a) is a process of sequentially extracting all the records having the same key field value to form a multi-type record string. Configure a field that is not a key field in the new record body that will be created.
In addition, the processing of b and c is processing for making the multi-type record string created in the temporary record string 2 into one record, and the created one record is the temporary record string 1
Will be linked to. The eighth processing example is different from the eighth processing example in that records having the same key field are extracted even if they are not consecutive, and the tenth processing example differs from each record of the temporary record string 2 in the key field. The difference is in removing. The temporary record string 1 becomes a new actual data part by executing the above-described processing until the elements of the actual data part of the search target data are exhausted. New search target data is created from the new data definition part generated as described above and the new actual data part.

FIG. 42 is an explanatory diagram of an example of the actual data part in the new search target data generated by the fourth hierarchical processing. In the fourth hierarchical processing described above, for example, the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes”. Consider the case. At this time, when the user inputs the hierarchization instruction shown in FIG. 31, the search target data created is a new search composed of the data definition part shown in FIG. 37 and the actual data part shown in FIG. Target data is generated. As shown in FIG. 42,
From the designated question record and opinion record, all the records having the same speaker field, for example, "a mountain a'kura" are sequentially extracted, and the speaker field "a mountain a '
"Kura", the speaker field is deleted from the extracted record, and a record type name with [-] added, for example, "question-: [...]" or "opinion-: [...]" A record having a record body composed of a composite field composed of a multi-type record string is created.

Also in the fourth example of the layering process described above,
It is also possible to construct a new actual data part only by a new record created from the specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is of a type not specified by the specifying means, the process returns to (1). Constitute. As a result, all records in which the value of the key field is the same as the key value are sequentially extracted from the actual data part of the read search target data, and the records are extracted by the key field and the field of the multi-type record column of the extracted record. The actual data part can be constructed by creating a main body and using the multi-type record string of the new record having only this new record.

FIG. 43 is an explanatory diagram of another example of the actual data part in the new search target data generated by the fourth hierarchical processing. As in the case of FIG. 42, for example, when the search target data having the data definition part shown in FIG. 32 and the actual data part shown in FIG. 33 is stored in the storage means 1 with the identifier “minutes” , Consider the case where the user performs the input shown in FIG. In this case, the actual data part of the created search target data is as shown in FIG. In FIG. 43, among the records shown in FIG. 42, question. A new actual data part is composed of only the opinion record.

Further, also in this fourth hierarchical example, the fields that can be designated as key fields are:
Only fields that are common to several types of records that are hierarchical. However, it can be configured so that a different key field can be specified for each record type. However, at this time, the atomic data type of the key field specified for each record type must be the same.

As a twelfth processing example, an example of performing non-hierarchical processing of records for a multi-type record string will be described. The non-hierarchical process of a record is a process of non-hierarchizing a non-hierarchical field, that is, one of composite fields in a record to be non-hierarchical to create a plurality of new records. Here, the record body of the new record is composed of fields constituting the record of each element of the non-hierarchical field and fields other than the non-hierarchical field. In the twelfth processing example, the generation unit 3 receives the identifier of the search target data, the record type name of the non-hierarchized record, and the designation of the non-hierarchized field from the designation unit 2. This information is
The user makes a designation through an input device which is a part of the designation means 2.

The generation means 3 receives the identifier of the data to be searched, the record type name of the non-hierarchized record, and the designation of the non-hierarchized field from the designation means 2, and then the identifier sent from the designation means 2. Data to be searched for is read from the storage means 1 into the main storage device. A new data definition part is created from the data definition part of the read search target data, and it is confirmed whether it satisfies the rules of the data definition part. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. After confirming that the rules of the data definition section are satisfied, a new actual data section is created from the actual data section of the retrieved search target data. New search target data is created from the new actual data section and the new data definition section. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

A method of creating a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the record definition part of the retrieved search target data.
A new record definition is added. For the record kind name of the new record definition, "." And the record kind name of the element B are connected to the record kind name of the record definition A from the element B of the non-hierarchical field of the data definition A having the specified record kind name. Created. Also, the record type of the new record definition is the record type of the record definition A with the field name of the non-hierarchical field removed, and the element B
It is created by concatenating the record types of the record definition that has as the record type name. A new record definition is created according to the created record type name and record type. This new record definition is created for each of all combinations of the data definition having the record type name designated by the designating means 2 and the element of the field name of the designated non-hierarchical field. The route definition part removes the specified record type name from the route definition part of the read target data, and connects the specified record type name and the field name element of the non-hierarchical field with ".". Add and create.

A method of creating a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following processes (1) and (2) are executed on the actual data part of the search target data. (1) The first record (referred to as R1 record) is extracted from the actual data part of the search target data. (2) If the record fetched in (1) is of a type not designated by the designating means, the record is linked to the end of the temporary record sequence 1 and the process returns to (1). If not, prepare the temporary record 2 which is the multi-type record string of the elements, and execute the following processing of a to i until the value of the non-hierarchical field of the fetched record has no elements. The record string 2 is connected to the end of the temporary record string 1, and the process returns to (1). I. The first record (referred to as R2 record) is extracted from the value of the non-hierarchical field of the R1 record. B. A new record type name is created by concatenating the record type name of the R1 record with the record type name of the R2 record. C. A new record body is created by concatenating the record body of the R2 record with the non-hierarchical field removed from the record body of the R1 record. D. The record consisting of the record type name created in (b) and the record body created in (c) is linked to the end of the temporary record string 2.

FIGS. 44 and 45 are explanatory views of an example of new search target data generated by the non-hierarchical processing.
In the non-hierarchical processing described above, for example, when the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. think of. At this time, the user specifies "minutes" as the identifier of the search target data, the question record as the type of the non-hierarchized record, and the {sent, answer} field as the non-hierarchical field, and instructs the non-hierarchicalization. When you enter, the search target data created will be
New search target data composed of the data definition part shown in FIG. 4 and the actual data part shown in FIG. 45 is generated.

As shown in FIG. 44, in the new data definition section, from the record type name “question” and the record type names “statement” and “answer” of the record of each element of the non-hierarchical field, “question / statement” is displayed. ] And “Question / Answer” record type names have been added. This added record definition record type is one element of the non-hierarchical field, for example, in the question / utterance record, the non-hierarchical field {utterance, answer} is deleted from the record definition record type of the question record. It is created by adding the record type of the utterance record, that is, "content".
The same applies to questions and answers. In addition, the new route definition part deletes the specified record type name “Question”, creates the “Question / Speak” from the specified record type name and the record type name of the record of the element of the non-hierarchical field, "Question / Answer" has been added.

As shown in FIG. 45, the actual data portion is composed of a record which is not designated and a new record which is created for each element of the non-hierarchical field from each designated question record. The record body of the created new record is, for example, in the case of the first comment record among the elements of the non-hierarchical field of the question record in which the speaker is “a mountain a'zo”, it is non-hierarchical from the question record. It is composed of the remaining field "a mountain a'zo" excluding the field and the record body "this invention is ..." Of the utterance record that is an element of the non-hierarchical field.

In the non-hierarchical processing described above, among the elements of the actual data part of the retrieval target data, a new actual data part is created by a record of a type not specified and a new record created from the record of the specified type. Although configured, it is possible to configure a new actual data part only with a new record created from a specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is of a type not specified by the specifying means, the process returns to (1). Constitute. This allows
Of the elements of the actual data part of the retrieved search target data,
Only the temporary record 2 that is the result of executing the above-mentioned processes a to d on the designated type of record is concatenated to the temporary record 1, whereby a new actual data part can be configured.

FIG. 46 is an explanatory diagram of another example of the actual data portion in the new search target data generated by the non-hierarchical processing. As in the case of FIG. 45, for example, the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG.
In the case of the non-hierarchical field, the user specifies "minutes" as the identifier of the search target data, the question record as the type of the non-hierarchical record, and the {sent, answer} field as the non-hierarchical field. Consider a case where the hierarchical processing is designated. In this case, the actual data part of the created search target data is as shown in FIG.
In FIG. 46, among the records shown in FIG. 45, the question. Statement record and question.
The answer record constitutes a new actual data part.

As a thirteenth processing example, an example of performing record expansion processing for a multi-type record string will be described. The record expansion process creates new data (hereinafter referred to as expansion data) from the record,
This is a process of adding a field whose value is the created data. The extended data is atomic data type data, a multi-type record string, or the like.
In the thirteenth processing example, the selection processing can be specified as the operation of creating the extension data from the record. Of course, it may be configured so that processing other than this can be designated. This selecti
The on process is a process for creating a new multi-type record string by removing records that do not meet the conditions from the element of the value of the specific composite field of the record.

In the thirteenth processing example, the generation means 3 is
From the designation means 2, the identifier of the search target data, the record type name of the record to be extended, the record type name after the record is extended, and the designation of the operation for creating the extended data from the extended record are received. These pieces of information are designated by the user via the input device which is a part of the designation means 2.

FIG. 47 is an explanatory diagram of an example of the format of input information when performing the extension process. For the user to create extended data from the identifier of the search target data, the record type name of the record to be extended, the record type name after the record is extended, and the extended record via the input device of the specifying means 2. It shows an overview of the information to be entered when specifying the operation. In FIG. 47, the “identifier” is a part that specifies the identifier of the search target data. The “record type name” is a part that specifies the record type name of the expanded record. The “extended record type name” is a part that specifies the record type name after the record is extended.
The "operation designation" is a portion for designating an operation for creating extended data from a record to be extended.

FIG. 48 is an explanatory diagram showing an example of the format of the input information of "operation designation". In FIG. 48, "target field designation" is a portion for designating a composite field in the expanded record. The value of the composite field specified here is the target of the selection process. The “record type name” is a part for specifying the record type name.
“Condition” is a part for designating a condition for a record of the type described before “by”. If the "by" and "condition" parts are omitted, all the records specified in the "record type name" part are selected. The method of specifying the field in the "condition specification" part is the same as the condition specification of the first processing example.

FIG. 49 is an explanatory diagram of an example of input information when the expansion process is performed. In FIG. 49, it is specified that the search target is the search target data having the identifier “minutes”. Further, the record to be expanded is a question record, the record type name after expansion of the question record is "question '", and the operation for creating the expanded data is It is specified that the selection process is for the value of the answer field. Also, this selectio
The record to be selected in the n process is the utterance record, and since the “by” and “condition” portions are omitted, all the utterance records are selected.

The generating means 3 receives from the specifying means 2 the identifier of the search target data, the record type name of the record to be expanded,
After receiving the record type name after the record is extended and the designation of the operation for creating the extended data from the extended record, the retrieval target data having the identifier sent from the designation means 2 is stored from the storage means 1. Read to main memory. A new data definition part is created from the data definition part of the read search target data, and it is confirmed whether it satisfies the rules of the data definition part. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. After confirming that the rules of the data definition section are satisfied, a new actual data section is created from the elements of the actual data section of the read target data. New search target data is created from the created new actual data part and the new data definition part. The new search target data is displayed on the display unit 4 as C.
Displayed on the RT display, the user can see the search results.

How to create a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the record definition part of the retrieved search target data.
It has the record type name specified in "Expanded record type name" and has the record type name specified in "Record type name ^", and is specified in "Target field specification" at the end of the record type of the record definition. It adds a new record definition with a record type that adds the field name of the specified field. The route definition part is a route definition part of the read search target data, in which the one designated by the "record type name" is removed and the one designated by the "extended record type name" is added.

A method of creating a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following processes (1) and (2) are executed until there are no elements in the actual data part of the read search target data. (1) The first record is extracted from the actual data part of the search target data. (2) If the fetched record is of a type that is not designated by the designation means 2, the fetched record is linked to the end of the temporary record string 1 and the process returns to (1). If this is not the case, a new field whose value is a multi-type record string created by subjecting the value of the composite field of the record specified by the specifying means 2 to the selection process is created, and this new field is added to the end of the record body. to add. Furthermore, the record type name is changed to the one specified in the "extended record type name". The record created by this processing is linked to the end of the temporary record string 1, and the process returns to (1). When there are no more elements in the actual data part, the temporary record string 1 constitutes a new actual data part.

50 and 51 are explanatory views of an example of new search target data generated by the expansion processing. In the above-described expansion processing, for example, consider a case where search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. . At this time, when the user inputs the extension instruction shown in FIG. 49, the search target data created is
New search target data composed of the data definition part shown in FIG. 50 and the actual data part shown in FIG. 51 is generated.

As shown in FIG. 50, a question 'record is added to the new data definition section. The record type of this question 'record is the same as the record type of the question record,
A field called {speech, answer} specified in "Target field specification" is added. Further, in the route definition part, "question" is changed to "question '". As shown in FIG. 51, the actual data section is composed of an unspecified opinion record and an expanded new question 'record. The record body of the created new question 'record is, for example, a multi-type record string of all of the statement records among the records of the elements of the statement field {statement, answer} of the question record, and this is added to the end of the field of the question record. It is created by adding.

In the above-mentioned extension processing, a new actual data part is constructed by the record of the unspecified type among the elements of the actual data part of the retrieval target data and the new record created from the specified type of record. However, it is also possible to configure a new actual data part only by a new record created from the specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is of a type not specified by the specifying means, the process returns to (1). Constitute. As a result, a record of a type that is not designated by the designation means is not linked to the temporary record 1, and a new actual data part can be constructed by removing these records.

FIG. 52 is an explanatory diagram of another example of the actual data part in the new search target data generated by the expansion processing. Similar to the case of FIG. 51, for example, when the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user specifies the extension processing shown in FIG. In this case, the actual data portion of the created search target data is as shown in FIG. In FIG. 52, of the records shown in FIG. 51, the question 'record constitutes a new actual data part.

As a fourteenth processing example, an example of performing record replacement processing for a multi-type record string will be described. With record replacement processing, new data (hereinafter referred to as replacement data) is created from the record,
This is a process of replacing a field whose value is the created data with a specific field of the record. The replacement data is atomic data type data, a multi-type record string, or the like. In the fourteenth processing example, similarly to the thirteenth processing example, the selection processing can be designated as the operation for creating the replacement data from the record. Of course, it may be configured so that processing other than this can be designated. This selectio
The n process is the same as the 13th process example.

In the fourteenth processing example, the generation means 3
From the designating means 2, the identifier of the search target data, the record type name of the record to be replaced, the field to be replaced,
Receives the record type name after the record is replaced and the specification of the operation to create replacement data from the replaced record. These pieces of information are designated by the user via the input device which is a part of the designation means 2.

FIG. 53 is an explanatory diagram showing an example of the format of input information when performing the replacement process. The user inputs the identifier of the search target data, the record type name of the record to be replaced, the field to be replaced, through the input device of the specifying unit 2.
It shows an outline of the information to be input when specifying the record type name after the record is replaced and the operation for creating replacement data from the replaced record. In FIG. 53, the “identifier” is a part that specifies the identifier of the search target data. The “record type name” is a part that specifies the record type name of the record to be replaced. The "replacement field designation" is a portion that designates a field to be replaced. The “record type name after replacement” is a part that specifies the record type name after the record is replaced.
The "operation designation" is a portion for designating an operation for creating replacement data from a record to be replaced. The contents described here are almost the same as those in the 13th processing example.
For example, it can be specified in the format as shown in FIG.

FIG. 54 is an explanatory diagram of an example of input information when performing the replacement process. In FIG. 54, it is specified that the search target is the search target data having the identifier “minutes”. Also, the record to be replaced is a question record, and it is specified that the field to be replaced is the {speech, answer} field. Further, the record type name after the question record is replaced is "question '", and the operation for creating the replacement data is the selection process for the value of the {speak, answer} field of the question record. Is specified. In addition, the record to be selected in this selection process is the utterance record, and “by”
And the "condition" are omitted, all speech records are selected.

From the designating means 2, the generating means 3 uses the identifier of the data to be searched, the record type name of the record to be replaced, the field to be replaced, the record type name after the record is replaced, and the record to be replaced. After receiving the designation of the operation for creating the replacement data, the retrieval target data having the identifier sent from the designation unit 2 is read from the storage unit 1 into the main storage device. A new data definition part is created from the data definition part of the read search target data, and it is confirmed whether it satisfies the rules of the data definition part. If the result of the confirmation is that the rule is not satisfied, a display command of a message of data definition incompatibility is sent to the display means 4, and the process ends. After confirming that the rules of the data definition section are satisfied, a new actual data section is created from the elements of the actual data section of the read target data. New search target data is created from the created new actual data part and the new data definition part. The new search target data is displayed on the CRT display by the display unit 4, and the user can see the search result.

How to create a new data definition part will be described.
The atomic data definition part is the same as the atomic data definition part of the read search target data. The record definition part is the record definition part of the retrieved search target data.
Of the record type fields of the record definition that have the record type name specified in "Record type name after replacement" and the record type name specified in "Record type name", select "Target" in "Operation specification". A new record definition having a record type in which the field name of the field specified in "field specification" is replaced with the field name of the field specified in "replacement field specification" is added.
The route definition part is the route definition part of the read search target data, which has the one specified by "record type name" removed and the one specified by "record type name after replacement" added.

A method of creating a new actual data part will be described. First, a temporary record string 1 which is a multi-type record string having no element is prepared. The following processes (1) and (2) are executed until there are no elements in the actual data part of the read search target data. (1) The first record is extracted from the actual data part of the search target data. (2) If the fetched record is of a type that is not designated by the designation means 2, the fetched record is linked to the end of the temporary record string 1 and the process returns to (1). Otherwise, in the record designated by the designating means 2, the field designated by the "replacement field designation" is set to the value of the composite field designated by the "target field designation" among the fields of the retrieved record.
Replace the multi-type record string created by performing the section process with a new field whose value is. further,
Change the record type name to the one specified in "Record type name after replacement". The record created by this process is
The temporary record string 1 is linked to the end and the process returns to (1). When there are no more elements in the actual data part, the temporary record string 1 constitutes a new actual data part.

55 and 56 are explanatory views of an example of new search target data generated by the replacement process. In the above-mentioned replacement process, for example, consider a case where search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. . At this time, when the user inputs the replacement instruction shown in FIG. 54, the data definition section shown in FIG.
New search target data composed of the actual data part shown in FIG. 56 is generated.

As shown in FIG. 55, a question 'record is added to the new data definition section. The record type of this question 'record is the same as the record type of the question record,
The field of {sentence, reply} which is "target field designation" is {sentence, reply} which is "replacement field name"
It has been replaced with a field called, but in the end it has the same record type as the question record. Further, in the route definition part, "question" is changed to "question '". As shown in FIG. 56, the actual data portion is composed of an unspecified opinion record and a replaced new question 'record. The record body of the created new question 'record is, for example, a set of all the record of the statement record among the records of the elements of the field {statement, answer} of the question record, and this set is set to {statement of {statement,
It is created by replacing the answer field. Therefore, the answer record has been removed.

In the above-described replacement processing, a new actual data part is constructed by the record of the unspecified type and the new record created from the record of the specified type among the elements of the actual data part of the retrieval target data. However, it is also possible to configure a new actual data part only by a new record created from the specified type of record. That is, in the process of (2) in the process of creating the new actual data part described above, if the record extracted in (1) is of a type not specified by the specifying means, the process returns to (1). Constitute. As a result, a record of a type that is not designated by the designation means is not linked to the temporary record 1, and a new actual data part can be constructed by removing these records.

FIG. 57 is an explanatory diagram of another example of the actual data part in the new search target data generated by the replacement process. Similar to the case of FIG. 56, for example, when the search target data having the data definition part shown in FIG. 2 and the actual data part shown in FIG. 3 is stored in the storage means 1 with the identifier “minutes”. , Consider the case where the user specifies the replacement process shown in FIG. 54. In this case, the actual data part of the created search target data is as shown in FIG. In FIG. 57, of the records shown in FIG. 56, the question 'record constitutes a new actual data part.

The record extension processing and replacement processing described as the above-mentioned thirteenth and fourteenth processing examples apply to the first to fourteenth processing for the multi-type record string in the composite field.
This is processing for enabling the processing performed by the processing example.
Therefore, in the thirteenth and fourteenth processing examples, various operations such as those described in the second to fourteenth processing examples can be specified in addition to the selection processing for the operation of creating the extended data and the replacement data. It may be configured as follows.

In the above description of each processing example, individual processing has been described. However, each of these processings can be configured to be able to be executed independently, or configured to perform some processing. You can also do it. Furthermore, a data search device capable of performing all of the first to fourteenth processing examples may be configured. When configured so that multiple processes can be performed, the result of the process is the search target data, so it can be configured so that it can be the target of another operation. It can also be done. At this time, the input from the designating means 2 is
It is possible to sequentially specify, or it is possible to make a specification of processing to be continuously performed in a file and read the file to perform processing collectively. Furthermore, the execution of the above-described processing can be designated from another program or the like.

The search target data used in the above processing example is
They all represented the logical structure of a single document.
However, it is easy to create search target data corresponding to a sequence of a plurality of documents. 58 and 59 are explanatory diagrams of a sixth example of search target data. For example, FIGS.
The search target data shown in (1) represents a series of minutes. In this way, by connecting the columns of each document into a multi-type record string, it is possible to create search target data corresponding to the columns of a plurality of documents.

[0159]

As is clear from the above description, according to the present invention, not only the conventional record set consisting of one type of record but also the multi-type record sequence having a sequential relationship consisting of a plurality of types of records. On the other hand, it is possible to obtain a data search device capable of performing various processes. In addition, by providing some functions of these processes, the user can operate a multi-type record string expressing logical information of a document by designating a combination of primitive operations that are easy to understand. In addition to the above, there is an effect that it becomes easy to understand what kind of processing the designated operation is, even in the processing for the multi-type record string.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of a data search device of the present invention.

FIG. 2 is an explanatory diagram of a data definition unit in a first example of search target data.

FIG. 3 is an explanatory diagram of an actual data part in a first example of search target data.

FIG. 4 is an explanatory diagram of an example of a format of input information when performing a selection process.

FIG. 5 is an explanatory diagram of an example of input information when performing a selection process.

FIG. 6 is an explanatory diagram of an example of an actual data part in new search target data generated by the selection process.

FIG. 7 is an explanatory diagram of another example of the actual data part in the new search target data generated by the selection process.

FIG. 8 is an explanatory diagram of a data definition unit in a second example of search target data.

FIG. 9 is an explanatory diagram of an actual data part in a second example of search target data.

FIG. 10 is an explanatory diagram of an example of a data definition unit in new search target data generated by the union process.

FIG. 11 is an explanatory diagram of an example of an actual data part in new search target data generated by the union process.

FIG. 12 is an explanatory diagram of a third example of the actual data part of the search target data.

FIG. 13 is an explanatory diagram of an example of a data definition unit in new search target data generated by the intersection process.

FIG. 14 is an explanatory diagram of an example of an actual data part in new search target data generated by the intersection process.

FIG. 15 is an explanatory diagram of an example of an actual data part of new search target data generated by the difference process.

FIG. 16 is an explanatory diagram of a data definition unit in a fourth example of search target data.

FIG. 17 is an explanatory diagram of an actual data part in a fourth example of search target data.

FIG. 18 is an explanatory diagram of an example of a data definition unit in new search target data generated by the cartesian product process.

FIG. 19 is an explanatory diagram of an example of an actual data part in new search target data generated by the cartesian product process.

FIG. 20 is an explanatory diagram of an example of a format of input information when performing a rearrangement process.

FIG. 21 is an explanatory diagram of an example of input information when performing a rearrangement process.

FIG. 22 is an explanatory diagram of an example of a data definition unit in new search target data generated by the rearrangement process.

FIG. 23 is an explanatory diagram of an example of an actual data part in new search target data generated by the rearrangement process.

FIG. 24 is an explanatory diagram of another example of the actual data part in the new search target data generated by the rearrangement processing.

FIG. 25 is an explanatory diagram of an example of a format of input information when performing extraction processing.

FIG. 26 is an explanatory diagram of an example of input information when performing extraction processing.

FIG. 27 is an explanatory diagram of an example of a data definition unit in new search target data generated by extraction processing.

FIG. 28 is an explanatory diagram of an example of an actual data portion in new search target data generated by the extraction processing.

FIG. 29 is an explanatory diagram of another example of the actual data part in the new search target data generated by the extraction processing.

FIG. 30 is an explanatory diagram of an example of a format of input information when performing a layering process.

FIG. 31 is an explanatory diagram of an example of input information when performing a hierarchical process.

FIG. 32 is an explanatory diagram of a data definition unit in a fifth example of search target data.

FIG. 33 is an explanatory diagram of an actual data part in a fifth example of search target data.

FIG. 34 is an explanatory diagram of an example of a data definition unit in new search target data generated by the first hierarchical processing.

FIG. 35 is an explanatory diagram of an example of an actual data part in new search target data generated by the first hierarchical processing.

FIG. 36 is an explanatory diagram of another example of the actual data part in the new search target data generated by the first hierarchical processing.

[Fig. 37] Fig. 37 is an explanatory diagram of an example of a data definition unit in new search target data generated by the second hierarchical processing.

FIG. 38 is an explanatory diagram of an example of an actual data part in new search target data generated by the second hierarchical processing.

FIG. 39 is an explanatory diagram of another example of the actual data part in the new search target data generated by the second hierarchical processing.

FIG. 40 is an explanatory diagram of an example of an actual data portion in new search target data generated by the third hierarchical processing.

FIG. 41 is an explanatory diagram of another example of the actual data part in the new search target data generated by the third hierarchical processing.

FIG. 42 is an explanatory diagram of an example of an actual data part in new search target data generated by the fourth hierarchical processing.

FIG. 43 is an explanatory diagram of another example of the actual data part in the new search target data generated by the fourth hierarchical processing.

[Fig. 44] Fig. 44 is an explanatory diagram of an example of a data definition unit in new search target data generated by non-hierarchical processing.

FIG. 45 is an explanatory diagram showing an example of an actual data part in new search target data generated by the non-hierarchical processing.

FIG. 46 is an explanatory diagram of another example of the actual data part in the new search target data generated by the non-hierarchical processing.

[Fig. 47] Fig. 47 is an explanatory diagram of an example of a format of input information when performing expansion processing.

FIG. 48 is an explanatory diagram of an example of a format of input information of “operation designation”.

[Fig. 49] Fig. 49 is an explanatory diagram of an example of input information when performing expansion processing.

FIG. 50 is an explanatory diagram of an example of a data definition unit in new search target data generated by the extension process.

[Fig. 51] Fig. 51 is an explanatory diagram of an example of an actual data portion in new search target data generated by expansion processing.

FIG. 52 is an explanatory diagram of another example of the actual data part in the new search target data generated by the extension processing.

FIG. 53 is an explanatory diagram showing an example of a format of input information when performing a replacement process.

FIG. 54 is an explanatory diagram of an example of input information when performing a replacement process.

FIG. 55 is an explanatory diagram of an example of a data definition unit in new search target data generated by the replacement process.

FIG. 56 is an explanatory diagram of an example of an actual data part in new search target data generated by the replacement process.

FIG. 57 is an explanatory diagram of another example of the actual data part in the new search target data generated by the replacement process.

FIG. 58 is an explanatory diagram of a data definition unit in a sixth example of search target data.

FIG. 59 is an explanatory diagram of an actual data part in a sixth example of search target data.

[Explanation of symbols]

1 storage means, 2 designation means, 3 generation means, 4 display means.

Claims (25)

[Claims] 1. A data search apparatus for extracting desired information from a record string having an order relation, a storage unit for storing a record string configurable by a plurality of types of records, and a record string stored in the storage unit. And a generation means for generating a new record string by processing the record string stored in the storage means on the basis of the processing request inputted from the specifying means. Data retrieval device characterized by. 2. The type of record from the designation means,
A condition for a record of the type is designated, and the generation unit is a record that is the type of the record designated by the designation unit from the first record string stored in the storage unit and does not satisfy the condition. The data search device according to claim 1, wherein a new second record string from which is deleted is generated. 3. The type of record from the designation means,
A condition for a record of the type is designated, and the generation unit is a record that is the type of the record designated by the designation unit from the first record string stored in the storage unit and does not satisfy the condition. 2. The data search device according to claim 1, wherein the second record string is created by removing the records not specified by the specifying means. 4. The generating means generates a new third record string which is a concatenation of the second record string after the first record string stored in the storing means. The data retrieval device described in. 5. The new generating means removes a record that does not exist in the second record sequence stored in the storage means from the first record sequence stored in the storage means. The data search device according to claim 1, wherein a record string is generated. 6. The generating means removes a record existing in the second record sequence stored in the storage means from the first record sequence stored in the storage means, and creates a new third record sequence. The data search device according to claim 1, wherein a record string is generated. 7. The generation unit is a field of a record included in a first record string stored in the storage unit, and a field of a record included in a second record string stored in the storage unit. , And a new third record string created for all combinations in the order of the records included in the first record string and in the order of the records included in the second record string. It produces | generates, The Claim 1 characterized by the above-mentioned.
The data retrieval device described in. 8. The type of record from the designation means,
The field arrangement order of the record of the type is designated, and the generation unit sets the field arrangement order of the record of the type designated by the designating unit of the first record string stored in the storage unit as described above. 2. The data search device according to claim 1, wherein a new second record string in which the arrangement order designated by the designation means is changed is generated. 9. The type of record from the designation means,
The field arrangement order of the record of the type is designated, and the generation unit sets the field arrangement order of the record of the type designated by the designating unit of the first record string stored in the storage unit as described above. 2. The data search device according to claim 1, wherein a new second record string is created by changing the order of arrangement designated by the designating means, and further removing a record of a type not designated by the designating means. . 10. The type of record and a part of a field group of the record of the type are designated by the designating unit, and the generating unit is configured to designate the first record string stored in the storing unit. A new second record string is generated by replacing the record of the type designated by the means with a new record consisting of only the fields designated by the designating means among the fields in the record. 1. The data search device described in 1. 11. The specifying unit specifies a record type and a part of a field group of the record of the type, and the generating unit specifies the first record string stored in the storage unit. The record of the type designated by the means is replaced with a new record consisting of only the fields designated by the designating means among the fields in the record, and the record of the type not designated by the designating means is removed. The data search device according to claim 1, wherein a new second record string is generated. 12. The type of record and a part of a field group of the record of the type are designated by the designating unit, and the generating unit selects one of the first record strings stored in the storing unit. A new second record string in which a subsequence of the type designated by the designating means and having the same field value of the field group designated by the designating means is replaced with a new record created by layering the substring. The data search device according to claim 1, wherein 13. The type of record and a part of the field group of the record of the type are designated by the designating unit, and the generating unit stores the record of the first record string stored in the storing unit. Of these, a subsequence of the type designated by the designating means and having the same field value of the field group designated by the designating means is replaced with a new record created by hierarchizing the subsequence. 2. The data search device according to claim 1, wherein a new second record string is created by removing a record of a type that is not designated by the designation means. 14. A new record created by hierarchizing the partial string is a record composed of a field group designated by the designating unit and a field having the partial string as a value. The data search device according to 12 or 13. 15. A new record created by hierarchizing the subsequence is a field group designated by the designating means, and each record in the subsequence is designated by the designating means in the field group of the record. 13. A record comprising a field having a value obtained by replacing a record consisting of unspecified one.
Alternatively, the data search device according to item 13. 16. The type of record and a part of a field group of the record of the type are designated by the designating unit, and the generating unit records in the first record string stored in the storing unit. Then, one or more records that are of the type designated by the designating means and have the same field value of the field group designated by the designating means are sequentially extracted and concatenated, and this record string is hierarchically created. A new second record string is created by concatenating the new record and the record of the first record string of the type not specified by the specifying means in the order of the records in the first record string. The data search device according to claim 1, wherein 17. The type of record and a part of the field group of the record of the type are designated by the designating unit, and the generating unit records in the first record string stored in the storing unit. Then, one or more records that are of the type designated by the designating means and have the same field value of the field group designated by the designating means are sequentially extracted and concatenated, and this record string is hierarchically created. The data retrieving apparatus according to claim 1, wherein the new record sequence is created by concatenating the new record sequence in the order of the records in the first record sequence. 18. The new record is a record composed of a field group designated by the designating means and a field having a value as a record string obtained by concatenating the one or more records in the order of extraction. Claim 1
The data search device according to 6 or 17. 19. The new record is obtained by removing the field designated by the designating unit from each record of the field group designated by the designating unit and a record string in which the one or more records are connected in the order of extraction. The data search apparatus according to claim 16 or 17, wherein the record is a record including a field having a record string as a value. 20. The type of record and a field that is a record string consisting of records of one or more types of field values are designated by the designating unit, and the generating unit is stored in the storing unit. Of the one record string, the record of the type designated by the designating means is designated, and each record of the record string which is the value of the field designated by the designating means of the fields of the record is designated in the record. 2. The data search device according to claim 1, wherein a new second record string is created by replacing the record string created by adding to a field group not specified by the means. 21. The type of record and a field, which is a record string consisting of records of one or more types of field values, are designated by the designating unit, and the generating unit is stored in the storing unit. Of the one record string, the record of the type designated by the designating means is designated, and each record of the record string which is the value of the field designated by the designating means of the fields of the record A new second record string is created by replacing with a record string created by adding to a field group not specified by the means, and further removing a record of a type not specified by the specifying means. Item 1. The data search device according to item 1. 22. The type of record and a creating method for creating new data from the type of record are specified by the specifying unit, and the generating unit is the first record stored in the storage unit. A new second record string is created by adding a field whose field value is the data created from the record according to the creating method specified by the specifying unit to the record of the type specified by the specifying unit of the column. The data search device according to claim 1, wherein 23. The designation unit designates a record type and a creation method for creating new data from the record of the type, and the creation unit creates a first record stored in the storage unit. A field having data created from the record according to the creating method specified by the specifying means as a field value is added to the record of the type specified by the specifying means of the column, and further specified by the specifying means. The data search device according to claim 1, wherein a new second record string is created by removing a record of a type that has not existed. 24. The type of record, a field of the record of the type, and a creating method of creating new data from the record of the type are specified by the specifying unit, and the generating unit stores in the storage unit. The field designated by the designating means of the record of the type designated by the designating means of the first record string, and the data created from the record according to the creating method designated by the designating means The data search device according to claim 1, wherein a new second record string replaced with a field to be a value is generated. 25. The specification unit specifies a record type, a field of the record of the type, and a creation method of creating new data from the record of the type, and the generation unit stores the data in the storage unit. The field designated by the designating means of the record of the type designated by the designating means of the first record string, and the data created from the record according to the creating method designated by the designating means The data search device according to claim 1, wherein a new second record string is generated by replacing a field as a value and further removing a record of a type not designated by the designation means.