JPH03282885A - Table retrieving device with learning function - Google Patents

Abstract

PURPOSE:To efficiently retrieve the record by saving retrieval key information related to the previous retrieval success and position information of the record of the retrieval success, as saved key information and saved record position information, comparing the retrieval key information and the saved key information at the time of a new retrieval request, and determining a retrieval range. CONSTITUTION:In response to a retrieval request, saved key information saved in a learning data saving means 11 and retrieval key information related to the present retrieval request are compared, and in the case they coincide with each other, saved record position information saved in the learning data saving means 11 is delivered to a retrieval result output means 15 and the processing is finished, and in the case of noncoincidence, based on a large/small relation of the retrieval key information and the saved key information and the saved record position information saved in the learning data saving means 11, a retrieval range being suitable for the present retrieval request and informed together with the retrieval key information to a sequential retrieving means 13. In this regard, in an initial state that significant information is not stored in the learning data saving means 11, the whole range of a table 2 to be retrieved is determined as a retrieval range. In such a way, the retrieval can be executed efficiently.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a device that performs a search on a sequential table in which records are arranged in descending or ascending order. The present invention relates to a table retrieval device with a learning function that efficiently performs table retrieval using the learning function.

[Conventional technology]

Conventionally, the following two types of methods are known as search methods for this type of table.

(1) The first method is a sequential search method in which the table is sequentially read from the first record and compared with the search key information related to the search request to determine the corresponding record. The result will be as shown in the figure.

(2) Divide the entire size of the second method table (total number of records) into two,
The two-part search method compares the record information at the boundary with the search key information, narrows the next search range by dividing into two based on the size relationship, and determines the corresponding record. The result will be as shown in Figure 5.

[Problem to be solved by the invention]

Conventionally, tables have been searched using the above-mentioned search methods, but each method has the following drawbacks.

■ In the first sequential search method, since the search always starts from the first record in the table, it takes a lot of time to search in situations where records near the end of the table are searched continuously.

■ The second two-part search method eliminates the drawbacks of the first sequential search method, but the number of records in the table is two.
If it is not a power of , it is necessary to perform division on odd records in one step of the two-part division process, which complicates the process.

■ In both the first sequential search method and the second two-part search method, the current search process proceeds regardless of past search processes, so in situations where adjacent records are searched consecutively, However, the search efficiency is the same as when records are searched randomly, and high-speed searches cannot be performed.

The present invention solves these conventional drawbacks.The purpose of the present invention is to solve complex problems in situations where adjacent records are searched continuously, even if the search location is near the end of the table. An object of the present invention is to provide a table search device with a learning function that can efficiently search without performing any processing.

[Means to solve the problem]

In order to achieve the above object, the present invention provides a table search device that searches a table in which a plurality of records are arranged in descending or ascending order, and a learning data storage means that stores storage key information and storage record position information. and comparing the storage key information stored in the learning data storage means in response to the search request with the search key information related to the current search request, and comparing the comparison result with the storage record stored in the learning data storage means. a search range determining means for determining a search range based on location information; and a search range determining means for sequentially reading records in the table within the search range determined by the search range determining means and comparing them with search key information related to the search request. It has a sequential search means, and a learning data updating means that updates the contents stored in the learning data storage means when the sequential search means succeeds in retrieval.

Further, in a preferred embodiment of the present invention, when a plurality of records are arranged in ascending order in the table, the search range determining means is configured to determine that when the search key information related to the current search request is smaller than the storage key information, The range from the beginning of the table to just before the position indicated by the stored record position information is determined as the search range, and if the range is large, the range from immediately after the position indicated by the stored record position information of the table to the tail of the table is determined as the search range. If the search key information related to the current search request is smaller than the storage key information, the search range determining means selects the search range from immediately after the position indicated by the storage record position information of the table. The range up to the end of the table is determined as the search range, and if the range is large, the range from the beginning of the table to just before the position indicated by the storage record position information is determined as the search range.

[Effect]

In the table search device with a learning function of the present invention, for example, when a record at a position in the table is searched in a search process based on search key information related to a previous search request,
The learning data storage means stores the search key information and the position information of the record as storage key information and storage record position information, and when a new search request is input, the search range determining means first stores the search key information and the position information of the record. The storage key information stored in the data storage means and the search key information related to the current search request are compared, and if they match, the record indicated by the storage record location information is the target record, and the following search process is performed. Although omitted, in the case of a mismatch, a search range is determined based on the comparison result and stored record position information stored in the learning data storage means, and the search means sequentially searches the table within the determined search range. The records therein are sequentially read out and compared with the search key information related to the search request. Then, when the sequential search means succeeds in retrieval, the learning data updating means updates the storage key information and storage record position information of the learning data storage means accordingly.

〔Example〕

Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. The table search device 1 with a learning function of this embodiment has a table to be searched 2 in which a plurality of fixed-length or variable-length records R1 to Rn are arranged in a predetermined order (descending order or ascending order).
This is a device that searches this searched table 2 in response to a search request and notifies the request source of the search results, and includes learning data storage means 11. Search range determining means 12. Sequential search means 13. It has learning data updating means 14 and search result outputting means 15. Each of the means 11 to 15 has the following functions.

・Learning data storage means 11 Holds save key information and save record position information.

- Search range determining means 12 Compares the storage key information stored in the learning data storage means 11 in response to the search request with the search key information related to the current search request, and if they match, the learning data storage means 11
The stored record location information stored in the search result output means 15 is passed to the search result output means 15 to complete the process, and if there is a mismatch, the magnitude relationship between the search key information and the stored key information and the stored record stored in the learning data storage means 11 are Based on the record position information, a search range suitable for the current search request is determined and sequentially notified to the search means 13 along with the search key information. Note that in an initial state where no relevant information is stored in the learning data storage means 11, the entire range of the searched table 2 is determined as the search range.

- Sequential search means 13 sequentially reads records within the search range determined by the search range determination means 12 from the searched table 2 and compares them with the search key information related to the search request, and when a matching record is found, that record is Location information search result output means 1
5 and also notifies the learning data updating means 14 of the record position information and search key information.

・Learning data update means 14 Search key information passed from the sequential search means 13.

The storage key information and storage record position information stored in the learning data storage means 11 are updated based on the record position information.

- Search result output means 15 Notifies the request source of the record in the searched table 2 indicated by the record position information passed from the search range determination means 12 or the sequential search means 13 as a search result.

Next, the operation of this embodiment configured as described above will be explained.

Now, in the searched table 2, there are multiple records R1~
Rn is arranged in ascending order, that is, from the smallest code to the largest record in the key information, and the learning data storage means 11
It is assumed that no useful information is stored in . Under such circumstances, if a search request is input using key information included in record R2 as the search key information, the search range determining means 12 will search sequentially because there is no useful information in the learning data storage means 11. The means 13 is notified of the entire range of the searched table 2 as the search range. In response to this, the sequential search means 13 sequentially reads records in the searched table 2, starting from the first record R1, for example, and compares them with the search key information related to the search request. Therefore, the sequential search means 13 notifies the search result output means 15 of the position information of the second record, and also notifies the learning data update means 14 of the record position information and search key information. Search result output means 15
acquires record R2 according to the notified record position information and notifies it to the request source as a search result. On the other hand, the learning data updating means 14 stores the search key information and record position information regarding the notified record R2 in the learning data storing means 11 as storage key information and storage record position information.

Next, for example, if a search request is input in which the key information included in record R3 is the search key information, the search range determining means 12 determines that the search key information is the learning data storage means 1.
Because it is larger than the save key information (key information of record R2) saved in 1, the save record position information (record R2 key information)
Record R3 immediately after the position indicated by (record position information)
The range from to the last record Rn of the searched table 2 is determined as the search range. As a result, the sequential search means 13 reads M[next record from the record R3 of the searched table 2 and compares it with the search key information.In this case, the search key information matches the key information in the record R3. , record R3 is immediately detected as the corresponding record, and is delivered to the request source by the search result output means 15. At this time, the learning data updating means 14 updates the storage key information and storage record position information in the learning data storage means 11 to the search key information and record position information regarding the record FR3.

In this way, in a situation where adjacent records are searched continuously, it is possible to quickly search for a target record by making use of the previous search results. In addition,
When a search request is made for the same record as the previous one, the search range determination means 12 detects this, and the search result output means 15 outputs the search results without sequentially operating the search means 13, resulting in a very fast search. becomes possible. Moreover, although the above-mentioned operation example is for a case where the records R1 to Rn are arranged in ascending order, it is similarly applicable to a case where the records are arranged in order of failure.

FIG. 2 is a block diagram showing an example of the hardware configuration of another embodiment of the present invention. The table retrieval device with learning function of this embodiment has a plurality of records R as explained in FIG.
A searched table storage memory 6 that stores searched tables 2 arranged in ascending order of 1-Rn, a learning data storage memory 3 that stores storage key information and storage record pointers, and an output such as a CRT that outputs search results. It is composed of devices 4 and a central processing section 5 connected to these devices and executing processing as shown in FIG. 3, for example. The operation of this embodiment will be explained below with reference to FIGS. 2 and 3.

When searching the searched table 2, when search key information is given as information necessary for the search from an input device, a program, etc. (not shown), the central processing unit 5 starts the process shown in FIG. First, the storage key information stored in the learning data storage memory 3 and the above search key information are compared in size (processing S1), and the result is determined (processing S2). If the two are equal, the process ends normally, and in this case, the record pointed to by the save record pointer saved in the learning data save memory 3 in the searched table 2 is the target record, and that record is then output to the output device 4. Ru.

If it is determined in process S2 that the search key information is smaller than the storage key information, the desired record exists between the first record R1 of the searched table 2 and the record pointed to by the storage record pointer. is the record pointer value pointing to the first record R1.
1" and "save record pointer 1" are respectively stored in the end record pointer (step S3), and the process advances to step S5.

In addition, if it is determined in process S2 that the search key information is larger than the storage key information, the desired record is stored in the searched table 2 from the record pointed to by the storage record pointer.
Since the last record Rn0 exists between the last record Rn0, "Save record pointer + 1" is stored in the search record pointer, and a record pointer value "N" pointing to the last record Rn is stored in the end record pointer. Processing S4), Processing S
Proceed to step 5.

In processes 35 to S8, a sequential search is performed on the search range determined in process S3 or S4 as follows. First, the record pointed to by the search record pointer is read from the searched table 2, its contents are compared with the search key information (process S5), and the result is determined (process 56). If the two are equal, the search is successful. Therefore, the process proceeds to step S9. If the search key information is larger, it means that there is no record in the searched table 2 that has the same key information as the search key information, so it is processed as a search error. If the search key information is smaller, the search record pointer is incremented by 1 to search for the next record (processing S7), and then it is determined whether the search range has ended by comparing this search record pointer with the end record pointer. If it is determined that the process is not completed (step S8), the process returns to step S5 and the above-described process is repeated.

When it is determined in step S8 that the search range has ended, this means that there is no record in the searched table 2 that has the same key information as the search key information, so it is treated as a search error.

When a record with key information equal to the search key information is detected in process S6, in order to make the current search result useful for the next search, the save key information and save record pointer in the learning data save memory 3 are updated to the current search result. The search key information is replaced with the search record pointer (processing S9, 5IO), and the process ends normally.

Note that when the process ends normally, the record pointed to by the search record pointer is the target record, and the record is then output from the output device 4.

The above example of operation was applied to a table in which records are arranged in ascending order, but it can be similarly applied to a table in which records are arranged in descending order.

〔Effect of the invention〕

As explained above, the table search device with learning function of the present invention stores the search key information related to the previous successful search and the position information of the record for which the search was successful as saved key information and saved record position information. When making a new search request,
The search key information and the save key information are compared, and based on the size relationship, it is determined which side of the table should be the search range based on the position indicated by the save record position information, and the search is performed sequentially code by code within the search range. In this way, in a situation where adjacent records are searched continuously, the desired record can be obtained efficiently regardless of the search location and without complicated processing. You can search for records.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing an example of the hardware configuration of another embodiment of the present invention; FIG. 3 is a flowchart showing an example of processing by the central processing unit 5; FIG. 4 is a flowchart of the processing of the conventional sequential search method, and FIG. 5 is a flowchart of the processing of the conventional two-part search method. In the figure, 1...Table search device with learning function 11...Learning data storage means 12...Search range determining means 13...Sequential search means 14...Learning data updating means 15...Search results Output means 2...Search table 3...Learning data storage memory 4...Output device 5...Central processing unit 6...Search table storage memory

Claims (3)

[Claims] (1) In a table search device that searches a table in which a plurality of records are arranged in a predetermined order, the learning data storage means stores storage key information and storage record position information, and the learning data is responsive to a search request. The storage key information stored in the storage means and the search key information related to the current search request are compared, and a search range is determined based on the comparison result and the storage record position information stored in the learning data storage means. a search range determination means; a sequential search means for sequentially reading records in the table within the search range determined by the search range determination means and comparing them with search key information related to the search request; 1. A table search device with a learning function, comprising: learning data updating means for updating the content stored in the learning data storage means when the search is successful. (2) A plurality of records are arranged in ascending order in the table, and when the search key information related to the current search request is smaller than the save key information, the search range determining means indicates the save record position information from the top of the table. 2. The table search with learning function according to claim 1, wherein the range immediately before the position is determined as the search range, and when the range is large, the range from immediately after the position indicated by the saved record position information to the tail of the table is determined as the search range. Device. (3) A plurality of records are arranged in descending order in the table, and when the search key information related to the current search request is smaller than the save key information, the search range determining means selects a position indicated by the save record position information in the table. 2. The learning function according to claim 1, wherein the range from immediately before to the tail of the table is determined as the search range, and when the range is large, the range from the head of the table to just before the position indicated by the saved record position information is determined as the search range. Table search device.