JPH0567155A - Information retrieving device - Google Patents

Abstract

PURPOSE:To convert attribute names registered in each DB (data base) and to execute information retrieval by allowing attribute names registered in DBs and attribute names of the standard retrieval system to correspond to each other. CONSTITUTION:When a retrieval conditional statement is inputted, a knowledge base storage device 5 is used to divide the retrieval conditional statement into a data base name, a retrieval language, an attribute name, and an attribute; and if the attribute name in the retrieval conditional statement coincides with an attribute name registered in data base systems 3 and 4 as the information retrieval object, a retrieval language corresponding to the information schema of data base systems 3 and 4 is used to convert the retrieval conditional statement in accordance with the attribute name and the attribute value, and the conversion result is used for retrieval of data base systems 3 and 4 through an input/output device.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information retrieving apparatus capable of retrieving a database group under a retrieval condition that can be expressed by a unified standard retrieval formula. More specifically, it has been constructed and developed individually up to now. In the future, it will be essential to utilize databases that have different information schemas in a complex and integrated manner to carry out one task. However, these databases can be represented by standardized search formulas. Virtually by using the knowledge about the contents registered in each database to the database that has the attribute name different from the attribute name that appears in the search condition or the database that does not have the attribute surface itself while making it possible to search by the condition The present invention relates to an information search device that enables information search as one database.

[0002]

2. Description of the Related Art Conventionally, in order to be able to search a database (hereinafter, abbreviated as DB) group having different information schemas by a search condition that can be expressed by a unified standard search formula, each database is individually searched. The data is read once using the database management system (DBMS) and stored again in the standard information schema, that is, DB conversion.

[0003]

The above-mentioned conventional method is inefficient and has problems in terms of cost and time.

Further, although the information schema can be converted, the semantic level of the attribute name is not consistent, so that the DB
There is a problem that it is often impossible to obtain information that matches the search condition due to the difference in the attribute names registered in.

The present invention has been made in view of the above,
The purpose is to associate the attribute name registered in each DB with the attribute name of the standard search formula, and
For each B, convert to the attribute name registered in the DB and execute information search, or derive information that matches the final search condition from the search results obtained from multiple related attribute names Another object is to provide an information retrieval device capable of inferring a default value corresponding to the attribute name from domain knowledge about the contents of each DB.

[0006]

In order to achieve the above object, the information retrieval apparatus of the present invention inputs a retrieval condition sentence and outputs information matching the retrieval condition sentence as a retrieval result to various database system groups. I / O device having access interface of the above, knowledge about attribute names registered in the database system group, knowledge about information schema / search language of various database system groups, domain knowledge about contents registered in various database systems. An information retrieval device having a knowledge base storage device for storing and a control device for controlling the whole, wherein when a retrieval condition sentence is input to the control device through the input / output device, the control device transmits the retrieval condition sentence to the knowledge base. After dividing into database name, search language, attribute name and attribute using the storage device, If the attribute name in the condition sentence matches the attribute name registered in the database system for which information is to be searched, the search condition sentence is calculated from the attribute name and the attribute value using a search language corresponding to the information schema of the database system. The gist of the present invention is to have a means for converting and using it for searching the database system via the input / output device.

[0007]

In the information retrieval apparatus of the present invention, when a retrieval condition sentence is input, the retrieval condition sentence is divided into a database name, a retrieval language, an attribute name, and an attribute by using a knowledge base storage device, and then the retrieval condition is If the attribute name in the sentence matches the attribute name registered in the database system of the information search target, the search condition sentence is converted from the attribute name and attribute value using a search language corresponding to the information schema of the database system. Then, it is used for searching the database system via the input / output device.

[0008]

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

FIG. 1 is a block diagram of an information retrieval apparatus according to an embodiment of the present invention. In the figure, 1 is an input / output device for inputting SQL as a standard search formula and outputting a search result and for accessing various DBs, 2 is a control device for controlling the whole, and 3 and 4 are DB systems having different information schemas. Reference numeral 5 stores knowledge of attribute names registered in each DB and relational knowledge between attribute names, domain knowledge about contents registered in each DB, knowledge about information schema and search language of each DB, and the like. It is a knowledge base storage device.

2 and 3 show a knowledge base storage device 5.
A class frame in which the attribute name registered in the virtual DB, that is, the attribute name appearing in the search condition sentence is the slot name, which is an example of the knowledge stored in the virtual DB, is prepared and registered in this virtual DB. A virtual DB instance having an attribute name and its attribute value definition as a slot value is prepared. Furthermore, by preparing an instance frame corresponding to each one of various existing DB groups and describing the attribute name and attribute name definition registered in the DB in the slot value of the instance, the attribute name of the virtual DB Express corresponding knowledge with. As a result, it is possible to flexibly deal with increase / decrease in DB and addition / change of attribute names. Also, some knowledge of the domain is prepared in the contents of various DBs.
For example, in the facility DB, the location,
There is knowledge about the device number and device type (Fig. 3, domain knowledge 1).

Next, the operation of the above information retrieval apparatus will be described with reference to the flowchart shown in FIG.

First, the search condition input to the information search apparatus is analyzed (step 110). Here, a case where an information search request is made in the DB relating to the communication device under the following search conditions will be described as an example.

[Standard search formula (SQL)] SELECT Device type FROM Device list WHERE Location = 'Shinjuku North Building' Manufacturer name = 'Hitachi' Therefore, the analysis result is: Attribute name → device type, location, manufacturer name ○ Desired search result → Attribute name "attribute type" attribute value ○ Search condition → Location = 'Shinjuku North Building' and manufacturer's name = 'Hitachi'.

Next, the search condition is changed using the knowledge base (step 120). In this example, since the attribute name “device type” is not registered in the DB (slot value is NIL in FIG. 2), the domain knowledge of the DB is applied to extract the attribute names “location” and “device number”. The device type will be inferred. Therefore, the attribute name is changed to the device number. In other words, the search condition is: Attribute name → device number, location, manufacturer name ○ Desired search result → Attribute value of attribute name "device number" ○ Search condition → Location = 'Shinjuku North Building' and manufacturer name = 'Hitachi' Note that this will be described in detail later.

Next, in order to match the information schema of the equipment DB, a search expression is synthesized in the language used by the database management system (DBMS) of the DB using the changed search conditions (step 130). For this, various information schema conversion knowledge is utilized.

Next, a search is executed, and a solution is generated from the search result using the knowledge base (steps 140 and 15).
0). At this time, in the case where the attribute name in the search condition is not registered in the DB and it is necessary to infer from the plurality of attribute names, that is, the search result of the attribute name “device number” → “123
8 ”and the attribute value“ Shinjuku North Building ”of the attribute name“ Location ”: The domain type 1 can be applied from the description in the search condition to infer the device type“ D70 ”.

Next, the device type "D70" is returned as a reply (step 160).

Next, the process of changing the search condition will be described in detail below with reference to FIG.

The contents of the virtual DB instance and the DB instance are compared, and the corresponding attribute name of the slot value (CAR
The presence / absence of a copy) (NIL if none) is determined (step 210).

If there is a corresponding attribute name, the attribute name is compared with the attribute value definition, and the following processing is performed (step 220).

I. Exact match of attribute name and attribute value definition → Search is possible as it is (step 230).

Ii. The attribute names are different, but the corresponding attribute names exist, and the attribute value definitions match (step 24
0) → Change to the corresponding attribute name and search (step 3)
10) (Example: FIG. 2B).

Iii. When the attribute names match but the attribute value definitions are different (step 250) → The attribute value definition information is written in the frame for storing the search history and the search is performed as it is (step 300) (example) : Fig. 2B).

Iv. When both the attribute name and the attribute value definition are different (step 260) → The attribute name and attribute value definition information are written in the frame for storing the search history, and the search is performed by changing to the corresponding attribute name. Yes (step 310)
(Example: c in FIG. 2).

If there is no corresponding attribute name, it is determined whether or not the attribute name can be changed to a plurality of attribute names, and the following processing is performed (step 270).

I. Inference is made from attribute values corresponding to a plurality of attribute names registered in the DB. Therefore, here, the attribute name is changed to a plurality of attribute names, and the plural attribute name list is set in the attribute name slot of the search history frame (step 280).

Ii. Inferring the attribute value corresponding to the attribute value by utilizing the default value (the attribute value has an initial value under a certain condition) which is one of the domain knowledge of the DB To do. Therefore, here, the value "absent" is set in the attribute name slot of the search history frame, and the attribute name is deleted from the search condition (step 290).

Next, the process of generating a solution from the search result will be described in detail below with reference to FIG. First, the following judgment is made based on the contents of the search history frame (step 410).

If the attribute name corresponding to the attribute name of the search condition exists, but the attribute value definitions are different, the conversion function described in the DB instance is activated and converted (step 420).

When the attribute name corresponding to the attribute name of the search condition does not exist and a plurality of related attribute names are searched, the value of the attribute name is used by utilizing the value of the search result and the domain knowledge of the DB. Is inferred (step 430).

Example: The domain knowledge 1 is utilized from the location and the device number to infer the device type (d in FIG. 2).

When the attribute name corresponding to the attribute name of the search condition does not exist, the default value which is the domain knowledge of the DB is utilized to infer the attribute value corresponding to the attribute name (step 440).

Example: The attribute name "state" is a DB related to equipment
Then, the initial value "good" is always established. Therefore, it can be inferred that the attribute value is “good” (e in FIG. 2).

[0034]

As described above, according to the present invention,
By applying this device to multiple DB groups with different information schemas, it becomes possible to search information under search conditions that can be expressed by a unified standard search formula, and the DB groups can be used organically to improve work efficiency. It has improved dramatically, and D
It is economically superior because the B assets can be effectively used.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an information search device according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of representation of knowledge stored in a knowledge base storage device used in the information search device of FIG.

FIG. 3 is a diagram showing a representation example of knowledge stored in a knowledge base storage device used in the information search device of FIG. 1.

FIG. 4 is a flowchart showing the operation of the information search device of FIG.

5 is a flowchart showing a search condition changing process of the information search device of FIG.

6 is a flowchart showing a solution generation process from a search result in the information search device of FIG.

[Explanation of symbols]

 1 input / output device 2 control device 3 database system 4 database system 5 knowledge base storage device

Claims (5)

[Claims] 1. An input / output device which inputs a search condition sentence, outputs information matching the search condition sentence as a search result, has an access interface to various database system groups, and an attribute registered in the database system group. An information search device having a knowledge base storage device that stores knowledge about names, knowledge about information schemas and search languages of various database systems, and domain knowledge about contents registered in various database systems, and a control device that controls the whole If the search condition sentence is input to the control device through the input / output device, the control device divides the search condition sentence into a database name, a search language, an attribute name, and an attribute using the knowledge base storage device. ,
When the attribute name in the search condition sentence matches the attribute name registered in the database system of the information search target,
Information retrieval characterized by having means for converting a retrieval condition sentence from the attribute name and attribute value using a retrieval language corresponding to the information schema of the database system and utilizing it for retrieval of the database system via an input / output device. apparatus. 2. When the attribute name in the search condition statement does not match the attribute name registered in the database system of the information search target, and an attribute name having the same meaning as the attribute name exists,
The search condition sentence is converted from the synonymous attribute name using a search language corresponding to the information schema of the database system,
2. The information search device according to claim 1, further comprising means used for searching the database system via the input / output device. 3. When the attribute name in the search condition statement does not match the attribute name registered in the database system of the information search target, the search is performed by changing to a plurality of attribute names, The information search apparatus according to claim 1, further comprising means for inferring a value corresponding to the attribute name. 4. When the attribute name in the search condition statement does not match the attribute name registered in the database system of the information search target, the value of the attribute name is inferred from the default value which is the knowledge of the domain in the database system. The information retrieving apparatus according to claim 1, further comprising means. 5. Even if the definition of the value obtained from the search result is different from the definition of the attribute value corresponding to the attribute name in the search condition sentence, it has means for enabling the conversion. Information retrieval device described.