JP2004013906A - Multiple database heterogeneity elimination searching method and device, and recording medium recorded with multiple database heterogeneity elimination searching program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a multiple heterogeneity elimination searching method and a device for searching data from a plurality of databases by eliminating the heterogeneity of the plurality of databases which differ at least even partly in data structure, DBMS, and data representation even without specifying corresponding relation and transformation rule for each search request, and a recording medium recorded with a multiple heterogeneity elimination searching program. <P>SOLUTION: By storing and managing data representation forms and transformation functions between the data representation forms, etc., respectively in an information resource dictionary 7a and a transformation function library 9, and using the data representation forms and the transformation functions between the data representation forms, an inquiry transforming unit 5 creates a search request sentence by transforming the representation form of search request data to the representation form of data used by a search subject database 21, and a database communication control unit 11 searches the search subject database 21 using the search request sentence. <P>COPYRIGHT: (C)2004,JPO

Description

The present invention eliminates the heterogeneity of a data structure, a database management system (hereinafter abbreviated as “DBMS”), and a plurality of databases (hereinafter abbreviated as “DB”) that differ in at least a part of a data expression, and TECHNICAL FIELD The present invention relates to a method and an apparatus for resolving multiple database dissimilarity for retrieving data from a database and a recording medium on which a program for resolving multiple database dissimilarity is recorded.

Companies have introduced and used many database systems for various tasks such as customer management and sales management. Initially, the database was used for simple data management purposes, such as recording, changing, and using customer and sales information, but recently, for example, it has been asked for information on what sells products by region. It is required to use a database to obtain management strategic information such as.

For such a purpose, information from one database alone may not be sufficient, and it is necessary to integrate and use information from a plurality of databases, for example, combining customer information and sales information. . However, since each database is operated independently, information having the same contents may be represented in different expression formats and different data structures.

マ ル チ Multi-databases have been researched and developed as a mechanism for integrating and using multiple databases. A multi-database system is a system called a global schema that predefines the virtual structure of the database used by the user so that the user can search the distributed database without being aware of the expression format and structure of individual information. is there.

た め In order to construct this global schema, it is necessary to associate data items with the same meaning. Further, even when there is no global schema, it is necessary to associate a search item given as a search request with data items of a plurality of databases.

Until now, in the correspondence between the global schema or the search item and the data item, a combination operation is always specified in a search sentence from a plurality of items for a single DB, and the search result is obtained as one item. It was possible. In addition, when a plurality of DBs using different DBMSs have a one-to-n correspondence between items, simple conversions that can be expressed by a DBMS SQL statement such as simple connection of character strings and four arithmetic operations Has been implemented. Furthermore, when the correspondence between the items corresponds to a single DB in one-to-n correspondence, by specifying the item corresponding to the conversion rule in the search request statement, it is possible to specify the item as the search condition (for example, Patent Document 1).
Gengo Suzuki, Hiroki Machihara, Mitsuru Kawashita, Fragment View-Heterogeneity Resolution Method without Global View in Multidatabase, IEICE Technical Report, Japan, IEICE, January 23, 1997 Vol. 96, No. 469, p.25-30

In the conventional method described above, it is necessary to specify the correspondence between data items and its conversion rule for each search request, and when specifying as a search item of a plurality of databases, it is possible to cope with complicated conversion. Therefore, there is a problem that it cannot be specified as a search condition item of a plurality of databases.

The present invention has been made in view of the above, and it is an object of the present invention to specify a plurality of data structures, DBMSs, and data expressions that are different even at least in part from each other without specifying a correspondence or a conversion rule for each search request. It is an object of the present invention to provide a multi-database dissimilarity search method and apparatus for resolving the heterogeneity of a database and searching for data from a plurality of databases, and a recording medium on which a multi-database dissimilarity search program is recorded.

The present invention according to claim 1 eliminates the heterogeneity of a plurality of databases that differ in at least a part of a data structure, a database management system, and a data expression, and eliminates the heterogeneity of a plurality of databases that retrieves data from the plurality of databases. A search method that defines a data item, a data expression format, a relationship between data expression formats, a relationship between a plurality of data expression formats and one data expression format, a conversion function between data expression formats, and a plurality of data expressions. Defines a conversion function from a format to a data representation format and a conversion function from a data representation format to a plurality of data representation formats, and uses the above defined relationship and conversion function to perform the same for a search item. A combination of a plurality of meaningful data items is obtained as a search item, and the plurality of data items are combined. And gist to search multiple databases by the combined search items.

According to the first aspect of the present invention, a relationship between data expression formats and a conversion function between data expression formats are defined, and the same meaning is used for a search item using the defined relationship and conversion function. A combination of a plurality of data items is acquired as a search item, and a plurality of databases are searched by the search item obtained by combining the plurality of data items. Therefore, one search is performed for a plurality of databases having different data structures and the like. Even when an item corresponds to a plurality of data items, a search can be accurately performed.

In addition, the present invention according to claim 2 eliminates the heterogeneity of a plurality of databases that differ in at least a part of a data structure, a database management system, and a data expression, and searches for data from the plurality of databases. And a data function, a data expression format, a relationship between data expression formats, and a relationship between a plurality of data expression formats and one data expression format, and a conversion function between data expression formats. It stores and manages a conversion function from a plurality of data expression formats to one data expression format and a conversion function from one data expression format to a plurality of data expression formats, receives a search request, and searches for a search item of the search request. Detecting a target database, as well as the data expression format and relation information between the expression formats and the expression Using a conversion function between expressions, the search request data expression format is converted to the data expression format used in the search target database, and a search request statement that can be processed by the search target database is created. The gist of the present invention is to search a search target database using the search request sentence, convert the search result into a structure corresponding to the search request, and output the structure.

According to the second aspect of the present invention, a data expression format and a conversion function between the data expression formats are stored and managed, and the data of the search request is expressed using the data expression format and the conversion function between the expression formats. To convert the format to the expression format of the data used in the search target database and create a search request sentence, and to search the search target database using this search request sentence, the data structure, database management system, data expression Even for a plurality of different databases, the heterogeneity can be eliminated, and data retrieval can be performed accurately.

Further, the present invention according to claim 3 eliminates the heterogeneity of a plurality of databases differing even in at least a part of a data structure, a database management system, and a data expression, and searches for data from the plurality of databases. And a data storage management means for storing and managing a data item, a data expression format, a relationship between the data expression formats, and a relationship between a plurality of data expression formats and one data expression format. Conversion function storage management means for storing and managing a conversion function between a plurality of data representation formats, a conversion function from a plurality of data representation formats to a single data representation format, and a conversion function from a single data representation format to a plurality of data representation formats; To detect the search target database of the search item of the search request, and Using the data expression format stored in the management means and the relation information between the expression formats and the conversion function between the expression formats stored in the conversion function storage management means, the expression format of the data of the search request is searched for in the database to be searched. Using a search request sentence creating means for creating a search request sentence that can be processed in the search target database by converting the data into an expression form of the data used in the search request sentence created by the search request sentence creating means A conversion unit that converts a search result of the search unit into a structure corresponding to the search request using a search unit that searches a search target database, and the data storage management unit and the conversion function storage management unit, and outputs the result. And the means.

According to the third aspect of the present invention, a data expression format and a conversion function between the data expression formats are stored and managed, and the data of the search request is expressed using the data expression format and the conversion function between the expression formats. To convert the format to the expression format of the data used in the search target database and create a search request sentence, and to search the search target database using this search request sentence, the data structure, database management system, data expression Even for a plurality of different databases, the heterogeneity can be eliminated, and data retrieval can be performed accurately.

The present invention according to claim 4 eliminates the heterogeneity of a plurality of databases that differ in at least a part of a data structure, a database management system, and a data expression, and eliminates the heterogeneity of a plurality of databases that retrieves data from the plurality of databases. A recording medium on which a search program is recorded, which defines a data item, a data expression format, a relationship between data expression formats, a relationship between a plurality of data expression formats and one data expression format, and a conversion function between data expression formats. , A conversion function from a plurality of data representation formats to a single data representation format and a conversion function from a single data representation format to a plurality of data representation formats, and searching using the defined relationship and the conversion function. Obtain as a search item a combination of multiple data items that have the same meaning for the item, And summarized in that a plurality database heterologous eliminating search program to search multiple databases by the search items by combining a plurality of data items recorded in a recording medium.

As described above, according to the present invention, a relationship between data representation formats and a conversion function between data representation formats are defined, and the same meaning is used for a search item using the defined relationship and conversion function. A combination of a plurality of data items is obtained as a search item, and a plurality of databases are searched by the search item obtained by combining the plurality of data items. Therefore, a plurality of databases having different DBMSs, data structures, data expression formats, and the like can be obtained. On the other hand, even when one search item corresponds to a plurality of data items, data items can be searched accurately only by defining data expression formats. Further, according to the present invention, a data expression format and a conversion function between data expression formats are stored and managed, and the data expression format of the search request is searched for using the data expression format and the conversion function between expression formats. A search request sentence is created by converting the data into the expression format of the data used in the database, and the search target database is searched using the search request sentence. Therefore, for a plurality of databases having different data structures, DBMS, and data expressions, Even so, it is possible to eliminate the heterogeneity and perform data search accurately.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a multiple database dissimilarity search apparatus according to an embodiment of the present invention. The multiple-database dissimilarity resolving and searching apparatus shown in the figure has the same meaning in an environment in which a plurality of databases exist, but eliminates differences in the expressions of data items that have different expressions depending on the database, and further reduces the differences in the data structure. The search is performed from a plurality of databases by solving the problem. The user 1, the application program 3, the query conversion unit 5, the information resource management unit 7 including the information resource dictionary 7a, the conversion function library 9, the database communication control unit 11, It has a search result integration unit 13, and a plurality of databases 21 are connected to the database communication control unit 11 via a network 23.

In the multi-database dissimilarity search apparatus shown in FIG. 1, first, the user 1 inputs, through the application program 3, a search item serving as a search result as a search request, a search item specified as a search condition, and a search condition value. This search request is passed from the application program 3 to the inquiry conversion unit 5. The query conversion unit 5 clarifies the location of the search item on the search target database 21 using the information of the information resource dictionary 7a managed by the information resource management unit 7, and determines the location of the search item according to the data structure of each database 21. Create a search request statement that can be processed in the searched database. At this time, when the expression format of the data of the search condition value is different between the search request and the search target database 21, the conversion function of the conversion function library 9 is used, and the data expression format is used in the search target database 21. Convert to format.

The search request sentence to the search target database 21 created by the query conversion unit 5 is sent to the search target database 21 connected via the network 23 via the database communication control unit 11, and the search processing is performed in each database 21. Be executed.

Then, the search result of each database is sent to the database communication control unit 11, and is sent to the search result integration unit 13 together with the search request passed from the application program 3.

The search result integration unit 13 uses the conversion function of the conversion function library 9 in accordance with the search request from the application program 3 and the information in the information resource dictionary 7a of the information resource management unit 7 to obtain a structure corresponding to the search request. Convert search results to expression format, expression format specified by the user. The search result converted by the search result integration unit 13 is returned to the application program 3 and output to the user 1.

In order to realize the above, in the present embodiment, the domain, which is the data representation format of the data item, is managed by the information resource dictionary 7a. Specifically, a local domain representing the data representation format of the data item of each database to be searched, a domain group composed of domains having the same meaning, a global domain representing the domain group, and a user among the domain groups Is a user domain that the user wants to output, a domain that the user will specify as a search condition, a composite domain created by combining and converting global domains of multiple domain groups, and a domain group that is an element of the composite domain Is the element domain.

In addition, as a function for converting each domain, a domain conversion function for converting between a global domain and each domain, a composite domain conversion function for generating a composite domain from a plurality of element domains, and generating one element domain from a composite domain The element domain conversion function to be performed is managed by the information resource dictionary 7a and is incorporated in the conversion function library 9.

Further, in order to manage the domain information, the information resource dictionary 7a stores a local domain and data items. For example, a data item management table as shown in FIG. Domain group management table as shown in FIG. 6 (a) for managing domain, global domain, and global domain, domain conversion management table as shown in FIG. 6 (b) for managing conversion functions between domains, composite domain and element domain A composite domain management table as shown in FIG. 6C for managing the conversion function to the element domain, a composite domain conversion management table as shown in FIG. It has a user domain management table as shown in FIG. Management information, to accumulate. Further, drain information is set in advance in the table in the information resource dictionary 7a, and necessary conversion functions are registered in the conversion function library 9.

Next, the operations of the query conversion unit 5 and the search result integration unit 13 will be described with reference to the flowcharts shown in FIGS.

First, the operation of the inquiry conversion unit 5 will be described with reference to the flowchart shown in FIG. The query conversion unit 5 receives the search request from the application program 3 and divides it for each search item to perform processing (step S11). First, for the search item, the “search item corresponding data item estimation” function clarifies the data item corresponding to each item and the location on the local database (step S13), and “estimated data item domain group”. The "search" function clarifies the estimated domain and domain group of the data item from the data item management table and the domain group management table (step S15). Subsequently, it is checked whether or not the domain group is a composite domain by using the composite domain management table (step S17). At this time, if it is a composite domain, the "element domain search" function lists the element domains from the composite domain management table (step S19). The data item corresponding to each element domain is clarified on the local database by the "element domain correspondence data item estimation" function (step S21).

It is checked whether or not the search item handled so far is the item specified as the search condition (step S23). If so, the data value specified as the search condition value is replaced with the "search condition value element". Using the "domain decomposition" function, the value expressed in the composite domain is converted into the element domain (step S25). Further, the data specified as the search condition value is converted into the local domain by using the "search condition value local domain conversion" function (step S27).

If the search item does not correspond to the composite domain (NO in step S17), a "domain group corresponding data item estimation" function is performed (step S29), and it is determined whether or not the search item is specified as a search condition. Check (step S31), and if so, create a search request using the "search condition value local domain conversion" function (step S33).

組 み 合 わ せ By combining the search items and search conditions obtained as a result, a search request sentence suitable for the database to be searched is generated.

The generated search request sentence is transmitted from the database communication control unit to each database 21 to be searched via the network 23, each database 21 is searched, and the database communication control unit 11 receives the search result again. The database communication control unit 11 passes the search request and the search result received from the application program 3 to the search result integration unit 13.

Next, the operation of the search result integrating unit 13 will be described with reference to the flowchart shown in FIG. Using the “local domain → global domain conversion” function of the search result obtained as described above, the value expressed in the local domain is converted into the global domain (step S51). Of the data of the search results, those corresponding to the search items generated from the composite domain are converted from a plurality of search result items into one search result item by using the "composite domain conversion" function (step S53). , S55). In the case of a domain group in which a user domain is set, the user is converted from the global domain to the user domain by the "user domain conversion" function (steps S57 and S59).

Similarly, in the case of a search result that is not generated from the composite domain, in the case of a domain group in which a user domain is set, the search result is converted into a user domain by the “user domain conversion” function (steps S61 and S63). Finally, the converted search result is returned to the application program.

Next, the operation of the multiple-database heterogeneity resolving and searching apparatus shown in FIG. 1 will be described in further detail with a specific example. In this specific example, FIG. 4 shows a database to be searched and data items and data managed by the database. In this example, two databases, a food DB that manages food sales and a clothing DB that manages clothing sales, are set as search target databases.

(5) As shown in FIGS. 5 and 6, database information and domain information are registered in the information resource dictionary 7a in advance. Further, names of domain conversion functions for performing conversion between defined domains and outlines of the processing are as follows.

・ Domain conversion function:
-Separation addition conversion: Converts YYYYMMDD format date to MM / DD / YYYY format by adding delimiter "/".-Separation deletion conversion: Converts delimiter from MM / DD / YYYY format date. To be converted to YYYYMMDD format ・ Heisei AD conversion: Convert two-digit year of Heisei to four-digit year ・ Heisei conversion: Convert four-digit year to 2-digit year of Heisei Domain conversion function:
-Year / month / day combination conversion: Combines 4-digit year display, 2-month month display, 2-day day display-Element domain conversion function:
・ Year / Year / Year Extraction: Extract YYYY which is a year from YYYYMMDD format ・ Year / Year / Month / Year Extraction: Extract MM which is month from YYYYMMDD format ・ Year / Year / Year Extraction: YYYYMMDD The DD which is the day is extracted from the date of the format. The search is performed in the environment set as described above. First, it is assumed that a search request for “product name and sales date of the product sold on January 1, 1998 is requested” from user 1 to application 3. In the application, this is described as “select product name, sales date, where =“ 01/01/1998 ””, and the search request is passed to the inquiry conversion unit 5. At this time, the description method from the application is arbitrary, and for example, only the search request item and the search condition may be specified.

Looking at the processing of the inquiry conversion unit 5 for each item, for the search item “product name”, the location of a database having “product name” as data is estimated from the information resource dictionary 7a. As a result, it can be seen that the "product name" is located in the "food DB" and the "clothing DB" as data items "product name", respectively. Subsequently, whether a domain is set for these two data items is searched from the data item management table of the information resource dictionary 7a. As a result, it is understood that no domain has been set.

Next, for the "sales date", which is a search item also specified as a search condition, the data item management table of the information resource dictionary 7a is used to create a database having "sales date" as data. Perform location estimation. As a result, it can be seen that “sales date” exists in the “food DB” as a data item “sales date”.

A search is made from the data item management table of the information resource dictionary 7a to determine whether a domain has been set for this data item. As a result, it can be seen that it is set as the “YYYYMMDD domain”. The domain information of this “YYYYMMDD domain” is searched from the domain group management table of the information resource dictionary 7a. As a result, it can be seen that the domain group is the "date domain group" and the global domain is the "YYYYMMDD domain". Whether the domain group is registered as a composite domain is searched from the composite domain management table of the information resource dictionary 7a. From this, it can be seen that the “date domain group” has been registered as a composite domain.

Next, an element domain group of the “year / month / day domain group” which is a composite domain is searched from the composite domain management table of the information resource dictionary 7a. It can be seen that the element domains are “year domain group”, “month domain group”, and “day domain group”. Domains belonging to each domain group are searched from the domain group management table of the information resource dictionary 7a. "Year domain group" includes [4-digit year, 2-digit year display], month domain group includes "2-digit month display", and day domain group includes "2-digit day display". You can see that there is. The column in which each domain is set is searched from the data item management table of the information resource dictionary 7a. Then, it can be seen that the data items corresponding to all three of “two-digit display in Heisei,” “two-digit month display” and “two-digit day display” belong to “clothing DB”.

The search condition value “01/01/1998” is converted according to the location of the search item “sales date” in the search target database. That is, the “sales date” data item corresponds to the food DB, and the “date domain” is a global domain, so no conversion is performed, but the “clothing DB” is input. "Sales date" corresponds to the data items "Year", "Month", and "Day", and the respective domains are "Heisei 2 digits display", "Month 2 digits display", "Day 2" Digit display ”, the search condition value“ 01/01/1998 ”is converted to the global domain“ 19980101 ”by the domain conversion function“ delimited deletion conversion ”, and then the element domain conversion function“ year / month / year ” It is decomposed into element domains “1998”, “01”, and “01” by “date extraction”, “date extraction”, and “date extraction”. Furthermore, since the local domain of the "year" data item is "Heisei 2 digits display", the global domain is converted to "H10", which is the local domain, by the domain conversion function "History conversion".

Finally, create a search sentence by combining each item. For the food DB, “select product name, sales date from sales where sales date =“ 19980101 ””. For the clothing DB, “select product name, year, month, day from purchase where year =“ H10 ”and month =“ 01 ”and day =“ 01 ””.

(4) This is transmitted from the database communication control unit 11 to each database 21 as a search request. Each database 21 searches each database using a corresponding search sentence. The search result is sent from the database communication control unit 11 to the search result integration unit 13.

In the search result integration unit 13, the search result of the food DB indicates that the set domain is the global domain and the operation related to the composite domain is not performed, but the “year of sales” belongs to the domain group “year”. Since the user domain of the “month / day domain group” is different from the global domain, it is output after being converted by the domain conversion function “separation addition conversion”.

(4) Since a search using the composite domain was performed on the clothing DB, the following processing is performed on the search result. First, in the database output result, the data value of “year” is converted into “4-digit year display”, which is a global domain, using a domain conversion function “Heisei Era conversion”. Subsequently, the converted data and the data values of “month” and “day” are converted into a “date / month / date domain group” that is a composite domain using a composite domain conversion function “date / month / date combination”. Since the result converted into the composite domain is different from the user domain as in the case of the “food DB”, the domain is converted into the user domain, and this result is output.

As a result, the search results returned to the application and the user are as follows.

FIG. 1 is a block diagram showing a configuration of a multiple database dissimilarity search apparatus according to an embodiment of the present invention. 3 is a flowchart illustrating the operation of a query conversion unit used in the multiple database heterogeneity resolving and searching device of FIG. 1. 2 is a flowchart illustrating the operation of a search result integration unit used in the multiple database heterogeneity eliminating search device of FIG. 1. FIG. 2 is a diagram illustrating an example of a search target database in the embodiment of FIG. 1. FIG. 5 is a diagram illustrating a data item management table defining data items for the search target database illustrated in FIG. 4. FIG. 5 is a diagram illustrating tables defining domains for the search target database illustrated in FIG. 4.

Explanation of reference numerals

5 Query Conversion Unit 7 Information Resource Management Unit 7a Information Resource Dictionary 9 Conversion Function Library 11 Database Communication Control Unit 13 Search Result Integration Unit 21 Database

Claims (4)

A data structure, a database management system, a multi-database heterogeneity resolving search method for resolving the heterogeneity of a plurality of databases differing even in at least a part of a data expression, and searching for data from the plurality of databases,
Define data items, data representation formats, relationships between data representation formats, and relationships between multiple data representation formats and one data representation format,
Define a conversion function between data representation formats, a conversion function from multiple data representation formats to one data representation format, and a conversion function from one data representation format to multiple data representation formats,
Using the defined relationship and the conversion function, a combination of a plurality of data items having the same meaning with respect to the search item is obtained as a search item, and a plurality of data items are obtained by combining the plurality of data items. A search method for resolving multiple database heterogeneity, characterized by searching a database. A data structure, a database management system, a multi-database heterogeneity resolving search method for resolving the heterogeneity of a plurality of databases differing in at least a part of a data expression, and searching for data from the plurality of databases,
Storing and managing data items, data expression formats, relationships between data expression formats, and relationships between a plurality of data expression formats and one data expression format;
Storing and managing a conversion function between data representation formats, a conversion function from a plurality of data representation formats to one data representation format, and a conversion function from one data representation format to a plurality of data representation formats;
A search request is received, a database to be searched for the search item of the search request is detected, and the data expression format of the search request is expressed using the data expression format and the relation information between the expression formats and the conversion function between the expression formats. Is converted to the representation format of the data used in the search target database, and a search request statement that can be processed in the search target database is created.
Using the created search request statement, search the search target database,
A search method for resolving disparities in multiple databases, wherein the search result is converted into a structure according to the search request and output. A data structure, a database management system, a multi-database heterogeneity resolving search device for resolving the heterogeneity of a plurality of databases differing in at least a part of a data expression, and searching for data from the plurality of databases,
Data storage management means for storing and managing data items, data expression formats, relationships between data expression formats, and relationships between a plurality of data expression formats and one data expression format;
Conversion function storage management means for storing and managing a conversion function between data representation formats, a conversion function from a plurality of data representation formats to one data representation format, and a conversion function from one data representation format to a plurality of data representation formats ,
A search request is received, a database to be searched for the search item of the search request is detected, and data expression formats and relation information between the expression formats stored in the data storage management unit and stored in the conversion function storage management unit. Creates a search request sentence that can be processed by the search target database by converting the data format of the search request to the data use format of the search target database using the conversion function between the expression formats Means for creating a search request sentence,
A search unit for searching a search target database using the search request sentence created by the search request sentence creation unit;
A conversion output unit that converts the search result of the search unit into a structure corresponding to the search request using the data storage management unit and the conversion function storage management unit, and outputs the converted structure. Heterogeneity resolution search device. A recording medium that records a data structure, a database management system, and a multiple database heterogeneity resolving search program for resolving the heterogeneity of a plurality of databases differing in at least a part of a data expression and searching for data from the plurality of databases. hand,
Define data items, data representation formats, relationships between data representation formats, and relationships between multiple data representation formats and one data representation format,
Define a conversion function between data representation formats, a conversion function from multiple data representation formats to one data representation format, and a conversion function from one data representation format to multiple data representation formats,
Using the defined relationship and the conversion function, a combination of a plurality of data items having the same meaning with respect to the search item is obtained as a search item, and a plurality of data items are obtained by combining the plurality of data items. A recording medium storing a search program for resolving multiple database heterogeneity, characterized by searching a database.

Images