JPH09319757A - Information retrieving system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily cope with the change or the addition of a retrieving engine and to output an integrated retrieving result by providing a common interface by common retrieval information so as to make the common interface correspond to plural individual APIs(application program interface). SOLUTION: A common interface part 11 decides computers 3 and 4 corresponding to data bases 31A and 31B designated as retrieving destinations and converts previously set common retrieving information into individual retrieving information corresponding to APIs 33A and 33B in data base systems 12A and 12B corresponding to the designated data bases 31A and 31B to give it to retrieving engines 32A and 32B of the pertinent data base systems 12A and 12B through APIs 33A and 33B. In addition, the common interface part 11 converts the retrieving result of an outputting form by the individual retrieving information obtained from the engines 32A and 32B through APIs 33A and 33B into an outputting form by means of common retrieving information and outputs it.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for searching information in a plurality of databases having different search interfaces, and more particularly to an information search system suitable for a network distributed environment in which processing is distributed to a plurality of systems.

[0002]

2. Description of the Related Art For example, when a database search system is constructed by using search software (search engine) for searching a full-text database, an API (Application Program Inte
rface: Application program interface)
It is necessary to build the system according to. That is, when the search engine is incorporated into the system to realize the search function of the database, the API unique to the search engine follows between the application program and the search engine.

Therefore, when changing the search engine, the application program itself must be changed corresponding to the search engine. Also, in the normal case, individual databases may be managed by different search engines. Therefore, when the application program searches a plurality of databases, it is necessary to use a plurality of individual APIs corresponding to the respective search engines.

[0004]

[Problems to be Solved by the Invention] Search Engine API
In each of the search engines, the function argument, the calling format, the search result output format, and the like are different. Therefore, as described above, it is necessary to change the interface of the application program depending on the search engine used. Therefore, when the search engine used is changed, the interface of the application program needs to be changed significantly.

Further, even when one application program uses a plurality of different search engines, it is necessary to use a plurality of APIs corresponding to the respective search engines.

In order to simplify the development of the application program, an attempt to make the interface between the application program and the search engine common is made, for example, in Japanese Patent Laid-Open No. 6-67867. However, in the method disclosed in this publication, the conversion from the common interface to each API is executed by the machine on the application side. Therefore, it is necessary for the application side to have a conversion program, and for the application side to execute API functions and the like. Therefore, there is a problem that the load on the application side is large.

Further, it is not easy to simultaneously search a plurality of databases managed by different search engines. Further, even if it is realized, it is extremely difficult to obtain the search result in a unified format because the output format of the search result is different for each search engine.

The present invention has been made in view of the above-mentioned circumstances, and can easily deal with a change or addition of a search engine and can output an integrated search result. The purpose is to provide.

Another object of the present invention is to provide an information search system that can easily search a plurality of databases.

[0010]

In order to achieve the above object, an information retrieval system according to the present invention corresponds to a plurality of databases and the plurality of databases, respectively.
A plurality of database search means for searching each database and the database search means corresponding to the database specified as the search destination are determined, and the search information in a format common to the plurality of database search means is specified. The database search means corresponding to the database is converted into individual format search information and given to the database search means, and the individual format search results obtained from the database search means are converted into common format search results. Common interface means for outputting
It is characterized by including.

In the information search system according to the present invention, a common interface for converting the search information and the search result is arranged so that a plurality of search means can be dealt with flexibly. Therefore, the change or addition of the search engine can be easily dealt with, and the integrated search result can be output.

Independent of the common interface means,
An application executing unit that executes an application that issues a database search request to the common interface unit and receives and processes a search result from the common interface unit may be further provided.

With this configuration, the application that issues the search request is independent of the common interface means, and the application itself does not need to execute a process for searching the database, for example, an API function. Therefore, the burden on the application is small. Normal applications can be used as they are.

It is also possible that a search instructing means for instructing a search of the database is arranged, and the common interface means interfaces between the search instructing means and the plurality of database searching means. In this configuration, the search instruction means is independent of the common interface means,
The search instruction means does not need to execute a process for database search, for example, an API function. Therefore, the load on the search instruction means can be reduced.

The application executing means and the common interface means are arranged on separate machines and connected by communication via standard input / output, and the database searching means corresponding to the respective databases are provided on the same machine. It may be placed at. With such a configuration, it is possible to efficiently disperse and reduce the load of the system in which the communication between the database and the database search means is large.

Further, a search instructing means for instructing a search of the database may be arranged, and the common interface means may interface between the search instructing means and the plurality of database searching means.

The common interface means stores the substantially same search information in the individual format and the search information correspondence table in which the correspondence relationship between the search information in the common format is registered, and the common search information in the common item. Based on the search item correspondence table, a database search means corresponding to the database of the search destination is converted into search information in an individual format, and a search instruction interpreting means to give to the database search means and a format unique to the database search means. The search result may be configured to convert the search result into the common format based on the search item correspondence table and output the converted result.

The search result generating means may include an integrating means for integrating the search results in the individual formats of the plurality of database search means and converting the search results in a common format.

When the database searching means searches the database and outputs with a score according to the degree of importance of the search result, the integrating means includes means for integrating the search results based on the score. May be included.

The search instruction interpreting means may generate synonym search information from the search information in the common format using a synonym dictionary.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. An information search system according to an embodiment of the present invention will be described with reference to FIG. FIG.
1 shows a configuration of an information retrieval system constructed in a network distributed environment in which processing is distributed to a plurality of computer systems.

The information retrieval system comprises an application program 10 for requesting retrieval of a database and processing retrieval results, a common interface section 11, a first database system 12A and a second database system 12B.

The application program 10 operates independently of the common interface unit 11, requests the common interface unit 11 to search the database, receives the search result from the common interface unit 11, and processes and processes the search result. The application program 10 is the first
Computer 1. 1 shows one application program 10 and the first computer 1, a plurality of computers on the network can independently execute the application program 10. The configuration of the application program 10 is also arbitrary.

The common interface section 11 is arranged independently of the application program 10. The interface with the application program 10 is standard input / output, and is designed to be easily operable on a machine different from the application program 10. Has been done. The common interface unit 11 includes a search instruction interpretation unit 21, a search result generation unit 22, a search item correspondence table storage unit 23, a database management table storage unit 24, and a thesaurus dictionary storage unit 25.

The first database system 12A has a first database 31A, a first search engine 32A and a first API 33A. The second database system 12B includes a second database 31B and a second database 31B.
It has a search engine 32B and a second API 33B.

The common interface unit 11, the first database system 12A and the second database system 12B are present on the second, third and fourth computers 2, 3 and 4, respectively.

The first database system 12A uses the first search engine 32A, which is a database search means, to generate the first database 3 based on the search request given to the first API 33A, which is a search interface.
1A is searched, and the search result is output via the first API 33A. Similarly, the second database system 12
B searches the second database 31B by the second search engine 32B based on the search request given to the second API 33B, and returns the search result to the second API 33B.
Output via B. Different search information is used for the search request and the search result in the first API 33A and the search request and the search result in the second API 33B.

For example, in a full-text search or the like, when it is necessary to acquire word frequency information and the like from a database, communication processing between the search engine and API and the database becomes heavy. Therefore, in the configuration of FIG.
A and 33B, search engines 32A and 32B, and databases 31A and 31B are arranged on the same machines 3 and 4, respectively, and the load of the search processing is efficiently distributed.

The common interface section 11 determines, by the application program 10, the computers 3 and 4 that construct the database systems 12A and 12B having the databases 31A and 31B designated as the search destinations. Further, it corresponds to search information (search information in a format common to the databases 31A and 31B) supplied from the application program 10 and APIs 33A and 33B of the database systems 12A and 12B corresponding to the specified databases 31A and 31B. The search engine 32A, 3 of the corresponding database system 12A, 12B
2B via API 33A, 33B.

Further, the common interface section 11 is a search engine 32 of the database systems 31A and 31B.
The search results in the individual output formats obtained from the A and 32B through the APIs 33A and 33B are converted into search information in a common format and output.

That is, the database management table storage section 24 stores the database identification information of each database 31A, 31B, for example, the database name and each database 3.
Stores a database management table in which the correspondence with the computers 3 and 4 in which 1A and 31B exist is registered.

The search item correspondence table storage unit 23 stores APIs 33A and 3B of the database engines 32A and 32B.
A search item correspondence table in which a correspondence relationship between individual search information for each 3B and common search information set in common for substantially the same information as these search information is registered is stored. The search information is, for example, a search item, a search word, a search logical expression, a search output item, and the like, and the correspondence relationship of the search information including at least the search item is registered in the search item table.

The search instruction interpreter 21 includes a database 31A designated by the application program 10,
Based on the database name of 31B, the corresponding computer 3 or 4 is determined by referring to the database management table. Further, the search instruction interpreting unit 21 refers to the search item correspondence table and sets common search information to the search engines 32A, 3A of the specified search destination databases 31A, 31B.
It is converted into individual search information for APIs 33A and 33B corresponding to 2B and given to the API 33A or 33B.

The search result generation unit 22 uses the database 31.
Individual search information that is searched by the search engines 32A and 32B of A and 31B and supplied via the APIs 33A and 33B, for example, the search result of the output format by the search output item is converted into the output format by the common search information. Output.

Next, the information retrieval system configured as shown in FIG. 1 will be described by further specific operation. The computer 1 has an application program 1
0 is executed and a database search request is issued based on the request of the application program 10. The database search request is, for example, common interface format tagged text information as shown in FIG. 2, and information such as a database name to be searched, a search item, a search word, an importance, a search logical expression, and a result output item. including. It is also possible to simultaneously request a search for a plurality of databases with one search request. The search instruction based on the search information in the common interface format is given to the search instruction interpretation unit 21 of the common interface unit 11 in the computer 2.

The search instruction interpretation unit 21 stores the database management table in the database management table storage unit 24 and stores the search item correspondence table in the search item correspondence table storage unit 23.

In the database management table, the database name designated by the search instruction is stored in which computer 3, 4
Information about which search engine 32A, 32B is located is registered. The search instruction interpreting unit 21 refers to this database management table and searches the search information from the application program 10 for each search engine 32A, 3A.
Distribute to 2B.

In the search item correspondence table, search information defined by the common interface or search information in which the search word, the degree of importance, the search logical expression, and the result output item are added to the search item, and each API 33A. , 33B, or the correspondence with the search information obtained by adding the search word, the degree of importance, the search logical expression, the result output item, and the like to the search item is registered.

The search instruction interpretation unit 21 refers to the search item correspondence table and converts the search information defined by the common interface into individual search information for each search engine of the databases 31A and 31B to be searched. That is, FIG.
As shown in, for the item name of "Title", by the database system, "TI
Different item names may be used, such as "TLE" or "SUBJECT". For this reason, the item name defined in the common interface, for example, "TI", is associated with the item name of the API for the search engine of each database in the search item correspondence table.

Further, the search instruction interpretation unit 21 responds to the input search information such as the search word by the thesaurus dictionary storage unit 2.
Using the thesaurus dictionary stored in 5 (that is, a synonym dictionary), search information such as a search word can be added.
By adding such synonym information, it is possible to achieve a more accurate search.

The search instructions distributed to the search engines 32A and 32B are the corresponding search engines 32A and 32B.
The search processing is performed on the databases 31A and 31B designated by the APIs (programs) 33A and 33B for use.

The search result obtained by the search is sent to the search result generating section 22 of the common interface section 11, is converted into the common interface format, and the search result is returned to the application program 10 of the computer 1.
The search result generation unit 22 also refers to the search item correspondence table and converts the search output item / format.

As described above, when the common interface unit 11 is used, a plurality of databases managed by different search engines or a plurality of databases managed by the same search engine are designated when designating a database to be searched. It is possible to It is possible to perform a search from all data in these designated databases, integrate the search results from each database into one, and output the search results.

Further, some search engines refer to the degree of importance designated in advance by a search instruction, give a score (that is, a score) to the searched data, and return the result sorted by this score. Then, when a plurality of different search engines each give such a score, the plurality of different score-giving search engines are searched at the same time, and the search results with the scores are scored by the search result generating unit 22. It is possible to sort and integrate according to the above, and output the result to the application program 10. In this case, API 33A, 3
By unifying the criteria for assigning the importance of 3B, it becomes easy to integrate the search results with the scores assigned to each of the databases 31A and 31B and return them to the application program 10.

For example, assume that a database 31A under the first search engine 32A and a database 31B under the second search engine 32B are searched at the same time. First, a search is performed for each database 31A, 31B, and a search result and a score for each database are calculated by each search API 33A, 33B. Next, the search result generator 22 integrates the respective search results in the order of scores. Simply outputting the search results for each database in order in the integration process does not have the advantage of outputting in order of importance. In this embodiment, as described above, the API 33
Since the scores of A and 33B are assigned with the same criterion, the search result generation unit 22 can integrate the scores in the order of the unified criterion. For example, 1 given in database A
Score of 00 and 10 given in database B
Since the score of 0 has the same degree of importance, the final integrated results are also sorted according to the degree of importance, and the user on the application program 10 side can also refer to the search results in order of importance.

The common interface used for database retrieval as described above can be utilized not only for retrieval but also for database management such as data registration, deletion or update in the database. The database system including the above-mentioned API and search engine is SQL.
It can also be replaced with a relational database system that allows inquiries by language.

By defining the common interface for identifying the database and converting the search information in this way, the search processing and the output of the search result are made common by flexibly supporting a plurality of APIs, and the search engine is shared. It is possible to easily deal with the change or addition of, and it is also possible to output the integrated search result. Therefore, the design and modification of the application program 10 are minimal, and it is possible to search a plurality of databases without being aware of the individual search engines. Moreover, the search interface can be unified,
Since the search can be performed in parallel by a plurality of search engines, the search time can be substantially shortened.

In order to disperse the addition of search processing, the search engine API is placed on the same machine as the search engine. In particular, in the case of assigning importance in full-text search, it is necessary to acquire word frequency information from the database, and communication processing between the API and the database becomes heavy. Therefore, it is more effective to distribute the API program. Can be dispersed.

[0049]

As described above, according to the present invention, it is possible to deal with a plurality of individual search means or search interfaces and easily deal with a change or addition of search means (search engine). It is also possible to output the integrated search result. Further, by making the application program (search instruction means) and the common interface independent, the load of the application program can be reduced, and its design is easy, and a general-purpose application can be used as it is. Become.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an information search system according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a database search request composed of text information in a common interface format for explaining the operation of the information search system of FIG.

FIG. 3 is a diagram showing an example of a search item correspondence table for explaining the operation of the information search system of FIG.

[Explanation of symbols]

1 to 4 computer (system) 10 application program 11 common interface unit 12A, 12B database system 21 search instruction interpretation unit 22 search result generation unit 23 search item correspondence table storage unit 24 database management table storage unit 25 thesaurus dictionary storage unit 31A, 31B Database 32A, 32B Search engine 33A, 33B API (application program interface)

Claims (9)

[Claims] 1. A plurality of databases, a plurality of database search means which respectively correspond to the plurality of databases and respectively search each database, and a database search means which corresponds to a database designated as a search destination are determined. The database search means converts the search information in the format common to the plurality of database search means into the search information in the individual format for the database search means corresponding to the specified database, and gives the converted information to the database search means. An information retrieval system comprising: a common interface unit for converting the retrieval result in the individual format obtained from the above into a retrieval result in the common format and outputting the retrieval result. 2. Independent of the common interface means,
2. The information search according to claim 1, further comprising: an application execution unit that executes an application that issues a database search request to the common interface unit, receives an search result from the common interface unit, and processes the received search result. system. 3. The application executing means is arranged on a first machine, and the common interface means is arranged on a second machine different from the first machine,
The application executing means and the common interface means are connected by communication via standard input / output, the database searching means corresponding to the databases are arranged on the same machine, and the interface means is connected via communication. The information retrieval system according to claim 2, wherein the information retrieval system is connected to. 4. A search instruction means for instructing a search of the database is further included, wherein the common interface means is means for interfacing between the search instruction means and the plurality of database search means. The information search system according to claim 1. 5. The common interface means stores a search item correspondence table in which a correspondence relationship between substantially identical search information in the individual format and search information in the common format is registered, and the common search information. , Based on the search item correspondence table, the database search means corresponding to the database of the search destination is converted into search information in an individual format, the search instruction interpretation means to give to the database search means, and the database search means Format search results,
The information search system according to claim 1, 2, 3, or 4, further comprising: a search result generating unit that converts the output into the common format and outputs the converted result based on the search item correspondence table. 6. The search result generation means includes integration means for integrating the search results in the individual formats of the plurality of database search means and converting the search results in a common format.
The information retrieval system according to claim 5, characterized in that. 7. The database search means searches the database, outputs with a score according to the importance of search results, and the integrating means integrates the search results based on the scores. The information retrieval system according to claim 6, further comprising: 8. The information search according to claim 5, wherein the search instruction interpretation means includes means for generating synonym search information from the search information in the common format by using a synonym dictionary. system. 9. The search instruction interpretation means refers to the database management table from the common search information, a database management table that registers a correspondence relationship between the identification information of each database and the location of each database search means. Then, the database to be searched is determined, the common search information is based on the search item correspondence table,
The information search system according to claim 5, further comprising: a unit that converts the search information in the individual format of the database search unit corresponding to the search destination database and supplies the search information to the database search unit. .