JP2002149648A - Method and device for integrated retrieval and storage medium with integrated retrieval program stored therein - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and device for integrated retrieval and a storage medium with an integrated retrieval program stored therein which realizes retrieval by a plurality of inquiry sentences having different retrieving items, data form of respective items and retrieving conditions with a plurality of logically and independently constructed tables as objects. SOLUTION: This device receives the a plurality of inquiry sentences from users, extracts all the retrieving items from the inquiry sentences, extracts the combination of identifiable retrieving items by referring to the synonym dictionary defining the synonym of a data item name, defines a data item in accordance with the combination of the identified retrieving items selected by the users, prepares the frame work of the table to be outputted as the result of retrieval, adds an insertion instruction to the respective inquiry sentences corresponding to the frame work of the table, corrects the order of the retrieval items, retrieves the table of the retrieval object by using the corrected inquiry sentence, and outputs the retrieved result to the user.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an integrated search method and apparatus and a storage medium storing an integrated search program, and more particularly, to a plurality of tables having different structures for a plurality of logically independent tables of a database. The present invention relates to an integrated search method and apparatus for obtaining a single search result using the query sentence described above, and a storage medium storing an integrated search program.

[0002]

2. Description of the Related Art A conventional database integrated search method is as follows.
A method is employed in which items are identified over a plurality of table boundaries using an item correspondence table prepared in advance, or a plurality of tables are joined according to records. By combining multiple tables by these methods, one large table is created,
By executing a search on the other hand, an integrated search is realized.

[0003] UNION in the database query language SQL is an operator for integrating search results obtained from a plurality of queries. Normally, when performing a search on multiple tables, each table is queried,
The use must sort out the obtained results, but as shown in FIG. 12, [list A] and [list B]
By using ON, the user can obtain a unified search result.

[0004] On the other hand, there is also a demand for performing a search on a plurality of tables that are logically independently constructed and obtaining an integrated search result. Therefore, there is a method of identifying synonymous data items included in each table and integrating those columns, and a method of integrating search item data related to the same record included in each table.

As such an integrated search technique, Japanese Patent Laid-Open No.
-891320 "Information search device". FIG. 13 shows a search by item identification using the item correspondence table. In this method, as shown in FIG. 13, an item correspondence table having the same name and the same meaning is automatically and interactively created in advance, and the corresponding items are identified and identified, so that the table exists over a plurality of tables. Merge columns. That is, an inquiry [SELECT name, age] having the same structure is performed for each information source, and the results are integrated.

Japanese Patent Application Laid-Open No. 11-175552, entitled "Database Searching Apparatus and Computer-Readable Recording Medium Recording Database Searching Program,"
As shown in (1), the table (name list, body measurement) to be searched is specified from the search item of the query sentence, and in each table, the data of the same record number (record number: 1) is [SEL
ECT name, age, height, weight]. The conditional search is a conditional search for the information after integration [WHERE
Age <30, weight <70]. As a result, results that cannot be obtained from individual tables [record number: 1, name:
AIUEO, age: 20, height: 170, weight: 60]
Search for.

[0007]

However, the above conventional method has the following problems.

For example, suppose that the inquiry shown in FIG. 16 is performed on a plurality of tables shown in FIG. 15 and the retrieval result shown in FIG. 17 is derived. In each table shown in FIG. 15 (diagnosis, appendicitis treatment, appendicitis patient), information on the patient is registered. The inquiry sentence (~) shown in FIG.
It is created in accordance with a request to "search for information on appendicitis patients".

Normally, when a plurality of search results are integrated using a UNION operator, the number of search items specified in each query sentence, the order, and the data type of each item need to match. However, since the number of items in each query sentence shown in FIG. 16 does not match, integration by UNION cannot be performed.

[0010] Further, in the above-mentioned JP-A-5-81320 and JP-A-11-175552, search conditions for a specific table are not considered. That is, since the search condition for a specific table cannot be specified, the meaning of the items or the table itself may be lost due to the integration of the tables. Therefore, FIG.
The search condition ("WHERE result = appendicitis") for the specific table to which the inquiry shown in FIG. 6 is specified becomes invalid. As shown in FIG. 18, when only the integration of data items is performed, data that does not satisfy the inquiry is included in the search result. If search conditions are specified for all tables as well as search items, data that satisfies the contents of the inquiry will be omitted from the search results. In each case, FIG.
Cannot be obtained. That is, FIG.
If only the merging of the data items in the item correspondence table shown in (1) ([SELECT name, age, and gender]) is performed, [diagnosis. Patient name, appendicitis treatment. Patient name, appendicitis patient. [Name] is changed to [Name] and [Appendicitis treatment. Age, appendicitis patient. "Age" is changed to "Age" and "Appendicitis patient. sex]
Are integrated into “gender”, respectively, and become search targets. As can be seen from the search result, the patient (normal) other than "appendicitis" is also searched from the item of "diagnosis" in the item correspondence table.

Next, when the search conditions ([SELECT name, age, gender], [WHERE result = appendicitis]) are applied to all the tables, the tables “Appendicitis treatment” and “Appendicitis patients” are obtained.
There is a problem that information is not retrieved.

Further, in Japanese Patent Application No. 5-83320, the correspondence definition is static, so that the table to be searched is changed, the name of a search item that can be input by the user as a search request is changed, or Even if the target table is the same, it is necessary to correct the correspondence definition each time the usage scene is changed. It is impossible to define every correspondence in advance. Further, the more definitions, the higher the possibility that the search result includes unnecessary information.

In a relational database, there is a concept called a view. A data search operation can be interpreted as creating a table having a new meaning from a base table (also called a table). The view defines this operation and the table as a virtual object having no substance. Since the view is the result of a search operation,
In order to perform a search on a plurality of tables having different data structures (also called schemas) and obtain an integrated search result, it is necessary to adjust a query sentence for each table. To make adjustments, it is necessary to be familiar with the structure of the table and the query method.

The present invention has been made in view of the above points, and is intended for a plurality of logically independently constructed tables.
It is an object of the present invention to provide an integrated search method and apparatus capable of searching by a plurality of query sentences having different numbers and orders of search items, data types of each item, and search conditions, and a storage medium storing an integrated search program.

A further object of the present invention is to provide an integrated search method and apparatus and an integrated search program capable of flexibly coping with a change of a search function such as a search item name that can be input as an information source or a search request, and a use scene of the search. Is to provide a storage medium in which is stored.

[0016]

FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention (claim 1) provides an integrated system for searching a plurality of logically independent tables of a database using a plurality of query statements having different structures and obtaining one search result. In the search method, a plurality of query sentences are received from the user (step 1), all search items are extracted from the query sentence (step 2), and a synonym dictionary defining synonyms of data item names is referred to. A combination of search items that can be identified is extracted (step 3), a data item is defined according to the combination of search items to be identified selected by the user, and a framework for a table to be output as a search result is created ( Step 4) In accordance with the framework of the table, an insert instruction is added to each query sentence to correct the order of the search items (Step 5), and the search is performed using the corrected query sentence. Find the table elephant (step 6), and outputs the result retrieved by the user (step 7).

According to the present invention (claim 2), a synonym dictionary is created as a dictionary that manages only the vocabulary of registered words without managing the correspondence between data items in the table to be searched.

According to the present invention (claim 3), when a table to be searched is searched, a search result is derived in one table format.

FIG. 2 is a diagram showing the principle of the present invention.

[0021] The present invention (claim 4) provides an integrated system for searching a plurality of logically independent tables of a database using a plurality of query statements having different structures and obtaining one search result. A search device, a synonym dictionary 180 in which synonyms are set for each registered word, a query sentence input unit 110 for receiving a plurality of query sentences from a user, and extraction of all search items from the query sentence Search item acquisition means 120, a search item identification determination means 130 that extracts a combination of search items that can be identified with reference to the synonym dictionary 180, and a search item identification An output table creation unit 140 that defines a data item according to the combination and creates a framework of a table to be output as a search result, and a table created by the output table creation unit 140 In accordance with the set, and adds the inserted instructions to each query statement, the query statement modifying means 150 for modifying the order of the search items, query statements modifying means 150
Table 19 to be searched using the query sentence corrected in
It has an inquiry execution unit 160 for searching for "0" and a search result output unit 170 for outputting the searched result to the user.

The present invention (claim 5) provides a synonym dictionary 180
A dictionary that manages only the vocabulary of registered words without managing the correspondence between data items in the table to be searched is used. The present invention (claim 6) provides an inquiry execution means 160
, The retrieval result is derived in one table format.

According to the present invention (claim 7), in the query sentence correcting means 150, for a query sentence in which the number, order, and data type of each search item are the same, the operator U
Includes means for making one inquiry sentence using NION.

The present invention (claim 8) provides an integrated system for searching a plurality of logically independent tables of a database using a plurality of query statements having different structures and obtaining one search result. A storage medium that stores a search program and defines a query sentence input process that accepts multiple query sentences from the user, a search item acquisition process that extracts all search items from the query sentence, and synonyms for data item names. With reference to the synonym dictionary, a data item is defined according to a search item identification determination process for extracting a combination of search items that can be identified, and a combination of search items to be identified that are selected by the user. According to the output table creation process that creates a table framework to be output as search results and the table framework created by the output table creation process, By adding the insertion instruction to have combined statement,
A query modification process that modifies the order of the search items, a query execution process that searches the search target table using the query modified by the query modification process, and a search that outputs the retrieved results to the user And a result output process.

The present invention (claim 9) further includes a process of creating a synonym dictionary as a dictionary that manages only the vocabulary of registered words without managing the correspondence between data items in the table to be searched. .

According to the present invention (claim 10), in the inquiry execution process, the retrieval result is derived in one table format.

According to the present invention (claim 11), in the query sentence correcting process, for a query sentence in which the number, order, and data type of each item are the same, the operator U
Includes the process of making one query using NION.

As described above, according to the present invention, a synonym dictionary is prepared, and when a plurality of query sentences having different structures are received, the query sentences are converted so that one search result is obtained. Into a single table, it is possible to easily derive integrated search results.

[0029]

FIG. 3 is a functional block diagram for explaining an outline of the present invention.

The integrated search apparatus 100 shown in FIG. 1 includes a query sentence input unit 110, a search item acquisition unit 120, a search item identification determination unit 130, an output table creation unit 140, a query sentence correction unit 150, and a query execution unit 160. Search result output unit 170, data item name meaning management dictionary 180, Table 1
90.

The data item name meaning management dictionary 180 manages data item names semantically using a synonym dictionary in which synonyms are set for each registered word. Hereinafter, the data item name meaning management dictionary 180 is referred to as a synonym dictionary.

The query sentence input unit 110 receives from the user 10 a plurality of query sentences having different numbers of search items, data types of respective items, and search conditions.

The search item obtaining unit 120 extracts all search items from each query sentence.

The search item identification determination unit 130 determines the identity of each search item extracted by the search item acquisition unit 120 with reference to the synonym dictionary 180. Items whose item names are defined as synonyms in the dictionary are regarded as identical.
However, even if the item names are synonymous, in practice, in order to prevent items having information with different meanings from being identified with each other, the user is presented with all combinations of items that can be identified with each other,
It is also possible to select a combination required as an output. Thereby, the combination of the identification of the identified search items is derived, and the selection of the identification is performed by the user.
Determine the data item name of the output table.

The output table creator 140 generates a command sentence for creating an output table based on the result of the identification by the search item identification determiner 130, and has all the data items extracted from the query. Create a table for output. In addition,
The form of the table includes, in addition to the base table, a view table (virtual table, only the definition for creating the data without the substance of the data) in the database.

The query sentence correcting unit 150 adds an insert command to each input query sentence so as to insert the search result into the output table, and corrects the order of the search items of the query sentence. Further, in consideration of improvement in search efficiency, a query sentence in which the number, order, and data type of each item are the same can be made into one query sentence using the operator UNION.

The query execution unit 160 executes a query to be corrected for each search target table, and sequentially inserts the search results into the output table. After inserting all results,
Completely duplicate lines can be deleted.

The search result output unit 170 outputs the output table into which the results of all the corrected inquiries have been inserted to the user as search results based on a plurality of query sentences.

A table 190 is a search target table constructed logically independently.

Next, the operation of the present invention will be described.

FIG. 4 is a flowchart for explaining the operation of the present invention.

Step 101) A synonym dictionary 180 is set as a precondition for performing the processing after step 102.
The synonym dictionary 180 defines synonyms of data item names, and does not define correspondence between items. A plurality of data items determined as synonyms by this dictionary are candidates for identification. Further, the user 10 can designate an item name registered in this dictionary and its synonyms as a search item. When a table to be searched is added and this table includes unregistered data items, or when the user 10 adds an item name that can be specified as a search item, the table is added to the synonym dictionary 180. Register the item name and synonyms newly. As the synonym dictionary 180, a dictionary set by the administrator of the search system or a commercially available thesaurus dictionary can be used.

Step 102) Inquiry sentence input unit 1
At 10, a plurality of inquiry messages are received from the user.

Step 103) Search item acquisition unit 120
, All search items are extracted from each inquiry sentence.

Step 104) The search item identification judgment unit 130 refers to the synonym dictionary, checks synonyms for all search items, and regards search items whose item names are synonyms as the same.

Step 105) Output table creation unit 140
, A table having all the search items as data items is created in consideration of the result of the identification.

Step 106) Inquiry sentence correcting unit 1
At 50, the order of the search items is corrected on the premise of output (or the structure of each search statement is corrected to be the same using NULL), and the search result is inserted into the output table. Modify as follows. Step 107) The query execution unit 160 executes each of the corrected search statements for each information source.

Step 108) Search result output section 170
Outputs an output table into which the results of all modified queries have been inserted as search results using a plurality of query statements.

As described above, it is possible to realize a search using a plurality of query sentences having different numbers of search items, data types of respective items, and search conditions.

[0050]

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

In the following embodiment, an example will be described in which a search is performed for the three information sources (diagnosis, appendicitis treatment, and appendicitis patients) shown in FIG. 15 based on the flowchart shown in FIG.

FIG. 5 shows an example of a synonym dictionary according to an embodiment of the present invention. In the figure, * indicates a backward match, for example, the patient's name is included in “* Name”.

In the following, a description will be given of the query sentence input processing after step 102.

Query sentence input processing (step 102)
To accept a plurality of inquiry messages from the user.
In the example, “appendicitis patient” is obtained from the three information sources shown in FIG.
It is assumed that three query sentences shown in FIG. 16 are input for the purpose of searching for. Note that, when each inquiry is individually performed, search results (to) shown in FIG. 6 are obtained.

Next, retrieval item acquisition processing (step 10)
As 3), all search items are identified from each query. In the example, 6
Items are identified as search items. [Diagnosis. Patient name, appendicitis treatment. Patient name, appendicitis treatment. Age, appendicitis patient. Name, appendicitis patient. Gender, appendicitis patient. Age] Next, as a search item identification determination process (step 104), a combination of search items that can be identified is identified by referring to the synonym dictionary 180. When the synonym dictionary 180 shown in FIG. 5 is referred to, it is possible to summarize at least three items.

[Patient name, patient name, name is “name”, age is “age”, and gender is “sex”] The data items of the table output as search results differ depending on the combination of search items to be identified. Therefore, all combinations are output as candidates so that the most suitable combination for the user's request can be selected. FIG. 7 shows all combinations of search items to be identified. In the figure, it means that items having the same number are identified with each other. All the combinations of the search items shown in FIG. 7 thus obtained are presented to the user,
Let the user choose.

Next, an output table creation process (step 10)
As 5), a data item is defined in accordance with a combination of search items identified by the user and identified, and a table framework to be output as a search result is created. This may be a unit that creates a table (including a virtual table) according to the combination of search items to be identified. Here, an example in which a table is actually created in a database will be described. FIG. 8 is a diagram for explaining the creation of a search result output table according to one embodiment of the present invention. In the case where search items are grouped into three items (FIG. Here is an example of the table creation command statement in (case).

As for the data type of each data item in the output table, the data type in the search target table 190 is applied as it is. For data items that have been identified, a data type that can absorb all data types of each search item before identification is applied. For example, when a search item having a character type of 20 characters and a search item having a character type of 15 characters are identified, the data item in the output table has a character type of 20 characters.

Next, a query sentence correcting process (step 1)
06), an insertion instruction is added to each input query so as to insert the search result into the output table,
Modify the order of search items. FIG. 9 shows the modified query sentences. FIG. 10 shows a query sentence for obtaining a result in the form of a virtual table, a view table.

Next, an inquiry execution process (step 10)
As 7), each query is performed on the search target table 190 using the query sentence into which the command shown in FIG. 9 is inserted. As a result, the output table is as shown in FIG. In order to improve search efficiency, at the stage of query execution, query sentences having the same number, order, and data type of search items are combined into one query sentence using the UNION operator, and the search is performed. I do.

After the full search is inserted, completely duplicated rows can be deleted.

Finally, search result output processing (step 10)
8) The output table (FIG. 17) is output to the user as a search result.

With the above processing, the present invention enables the following.

Conventionally, a search for obtaining one table as a result from a plurality of query sentences having different structures cannot be realized.
In the present invention, first, a table for outputting search results is created, and a plurality of query results are sequentially inserted therein,
By realizing a search using multiple query sentences with different numbers of search items, data types of each item, and search conditions, the search results are always output as one table, and it takes time for the user to fulfill the search results. The burden can be reduced.

Conventionally, in order to change a table to be searched, it is necessary to add an item name and define a correspondence with a registered item. If you change it, the setting related to data item name management only needs to be added and registered in the same way as unsupported data item name.
Flexible response to changes in the table to be searched.

Further, conventionally, an item correspondence table is prepared in advance, and only static integration according to a defined correspondence is performed. When a user request is changed, the item correspondence table is also changed accordingly. Although it is necessary to change, in the present invention, all search items are identified from a plurality of query sentences, and all combinations of items that can be identified by referring to the synonym dictionary are all output, so that selection is possible. Thus, it is possible to always integrate search results requested by the user. For example, in the above embodiment, it is assumed that the age of the information source "appendicitis treatment" refers to the age at the time of treatment, and the age of "appendicitis patient" refers to the current age. In this case, these data items should not be identified.
In the present invention, in selecting a combination of item identification, by selecting a combination that does not identify these items, FIG.
Can be obtained. In the conventional search method using the item correspondence table, in order to make such a change, the correspondence table must be corrected.

Although the above embodiment has been described based on the flowchart shown in FIG. 4, the flowchart is constructed as a program, and a disk device connected to a computer used as an integrated search device, a floppy (registered) The present invention can be easily realized by storing it in a portable storage medium such as a trademark (trademark) disk, a CD-ROM, or the like and installing it when implementing the present invention.

It should be noted that the present invention is not limited to the above-described embodiment, and various modifications and applications are possible within the scope of the claims.

[0069]

As described above, according to the present invention, it is possible to satisfy a request of a user who does not know the structure of a table or the like by performing a search using a plurality of query sentences having different structures as automatically generated. Become.

Further, the search function can flexibly respond to changes in information sources, search item names that can be input as search requests, search use scenes, and the like.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention.

FIG. 2 is a principle configuration diagram of the present invention.

FIG. 3 is a functional block diagram illustrating an outline of the present invention.

FIG. 4 is a flowchart for explaining the operation of the present invention.

FIG. 5 is an example of a synonym dictionary according to an embodiment of the present invention.

FIG. 6 is a diagram showing a result of executing each inquiry according to the embodiment of the present invention.

FIG. 7 is an example of a combination of data item identification according to an embodiment of the present invention.

FIG. 8 is a diagram for explaining creation of a search result output table according to one embodiment of the present invention;

FIG. 9 is an example of performing a search and inserting a result into an output table according to an embodiment of the present invention.

FIG. 10 is an example of a query sentence for obtaining a search result as a view table according to an embodiment of the present invention.

FIG. 11 shows a search result obtained by changing a combination of data items being identified in one embodiment of the present invention.

FIG. 12 is a diagram for explaining query sentence combination and integrated search by UNION.

FIG. 13 is a diagram for describing a search by item identification using an item correspondence table.

FIG. 14 is a diagram illustrating a search by data integration according to a record number.

FIG. 15 is an example of a search target information source.

FIG. 16 is an example of a query sentence having a different structure.

FIG. 17 shows a target search result.

FIG. 18 is a diagram for explaining a problem that a condition cannot be specified for a specific table.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 User 100 Integrated search apparatus 110 Inquiry sentence input means, Inquiry sentence input part 120 Search item acquisition means, search item acquisition part 130 Search item identification determination means, Search item identification determination part 140 Output table preparation means, Output Table creation unit 150 Query sentence modification unit, Query sentence modification unit 160 Query execution unit, Query execution unit 170 Search result output unit, Search result output unit 180 Synonym dictionary (data item name meaning management dictionary) 190 Table to be searched

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Yuichi Iizuka 2-3-1 Otemachi, Chiyoda-ku, Tokyo Nippon Telegraph and Telephone Corporation F-term (reference) 5B075 ND20 NK35 PQ02 QP03

Claims (11)

[Claims] 1. An integrated search method for searching a plurality of logically independent tables of a database using a plurality of query statements having different structures and obtaining one search result. Accepting a plurality of query sentences, extracting all search items from the query sentences, referring to a synonym dictionary defining synonyms of data item names, extracting combinations of search items that can be identified, According to the combination of search items to be identified by the user, data items are defined, a table framework to be output as a search result is created, and an insertion instruction is added to each of the query statements according to the table framework. And correcting the order of the search items, searching the table to be searched using the corrected query, and outputting the searched result to the user. If the search method. 2. The integrated search method according to claim 1, wherein the synonym dictionary is created as a dictionary that manages only the vocabulary of registered words without managing the correspondence between data items in the table to be searched. 3. The integrated search method according to claim 1, wherein a search result is derived in one table format when searching the search target table. 4. An integrated search apparatus for searching a plurality of logically independent tables of a database using a plurality of query statements having different structures and obtaining one search result, A synonym dictionary in which synonyms are set for registered words, a query sentence input unit for receiving a plurality of query sentences from a user, a search item acquisition unit for extracting all search items from the query sentence, With reference to the synonym dictionary, a search item identification determination means for extracting a combination of search items that can be identified with each other, and defining a data item according to the combination of search items to be identified that are selected by the user, Output table creating means for creating a framework of a table to be output as a search result; and inserting into each of the query sentences according to the framework of the table created by the output table creating means. With an input command
Query sentence correcting means for correcting the order of the search items; query execution means for searching a search target table using the query sentence corrected by the query sentence correcting means; and outputting the searched result to the user. And a search result output means for performing the search. 5. The integrated search device according to claim 4, wherein the synonym dictionary manages only the vocabulary of registered words without managing the correspondence between data items in a table to be searched. 6. The integrated search apparatus according to claim 4, wherein said query execution means derives the search result in one table format. 7. The query sentence correcting means includes means for making a single query sentence using an operator UNION for a query sentence in which the number, order, and data type of each item are the same. 4. The integrated search device according to 4. 8. A storage medium storing an integrated search program for searching a plurality of logically independent tables of a database using a plurality of query statements having different structures and obtaining one search result. A query sentence input process for receiving a plurality of query sentences from a user, a search item acquisition process for extracting all search items from the query sentence, and a synonym dictionary defining synonyms for data item names. Then, a data item is defined in accordance with a search item identification determination process for extracting a combination of search items that can be identified, and a combination of search items to be identified that are selected by the user, and is output as a search result. An output table creation process for creating a table framework, and according to the table framework created by the output table creation process, A query sentence correction process for adding an insert instruction to the query sentence to correct the order of the search items; anda query execution process for searching the search target table using the query sentence corrected in the query sentence correction process, A search result output process for outputting a search result to the user. A storage medium storing an integrated search program. 9. The method according to claim 8, further comprising the step of creating the synonym dictionary as a dictionary for managing only the vocabulary of registered words without managing the correspondence between data items in the table to be searched. A storage medium storing an integrated search program. 10. The storage medium storing the integrated search program according to claim 8, wherein the inquiry execution process derives a search result in one tabular form. 11. The query sentence correcting process includes a process of making a query sentence using the operator UNION for a query sentence having the same number, order, and data type of each search item. A storage medium storing the integrated search program according to claim 8.