JP3438805B2 - Database heterogeneity resolution search device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database heterogeneity resolving search device which has the same meaning in an environment in which a plurality of databases exist, but searches for data items having different expressions depending on the databases by eliminating the difference in the expressions. Regarding

[0002]

2. Description of the Related Art Companies have introduced and used many database systems for various operations such as customer management and sales management. Initially, the database was used for the purpose of simple data management such as recording, changing, and using information on customers and sales, but recently, information such as what are the selling points of products by region is requested. It is required to use the database to obtain information on management strategies such as.

For such purposes, there are cases where the information in one database alone is insufficient, and the information in a plurality of databases is integrated and used, for example, customer information and sales information are combined. There is a need. However, since each database is operated independently, the same content information may be expressed in different expression formats. When integrating information from multiple databases, it is necessary to eliminate the difference.

Until now, multi-databases have been researched and developed as a mechanism for integrating and using a plurality of databases. The multi-database system is a system that allows users to search distributed databases without being aware of the expression format and structure of individual information by predefining a virtual structure of a database used by users called a global schema. Is.

However, in this method, the user can see the data only in the format of this global schema, and one way is forced.

[0006]

[Problems to be Solved by the Invention]
In order to define a global schema, we needed to know the representation of data in all databases. However, it needs to be familiar with the contents of all databases, which is actually very difficult. Therefore, it is necessary to have a mechanism that can eliminate the difference in the representation of data without knowing as much as possible the representation format other than that of the own database.

[0007] Further, the user is required to obtain information of another database in the same expression format as the database he has. This is because I want to use the data in another database with the application I am using, but that application is made to match the format of my database. If this request is to be realized by an existing multi-database, it is inefficient because it is necessary to define a global schema having the same format as the user's database for each user.
Therefore, it is necessary to eliminate the difference in the representation of the data by defining the difference in the data piece by piece without defining the global schema. And there is a need for a mechanism that allows the user to easily change the expression format he wants to use.

The present invention has been made in view of the above,
The purpose is to provide a database heterogeneity resolving search device capable of freely searching each database in the expression format of each user, even if the expression of the data item in each database is different despite having the same meaning. To provide.

[0009]

In order to achieve the above object, the present invention according to claim 1 represents a data item having the same meaning in an environment in which a plurality of databases exist, but different expressions depending on the databases. This is a database heterogeneity resolving search device that resolves the difference between the above and the same meaning as the local domain management table that manages each database as the local domain, which is the representation format of the data item of each database. A domain group management table that manages each of the domains as a domain group, and selects and manages a global domain that is typically used as a representative in each domain group .
Both between the local domain and the local domain
A conversion function management table that manages conversion functions that perform unidirectional or one-way conversion, and a user
User domain that manages the user domain as a user domain
Main management table, along with searching for a local domain from the local domain management table corresponding to the database and the search data search data query sentence input by the user is stored, and domain group that contains the local domain The user
Domain search means for searching the user domain from the user domain management table , and a query to the database.
When the conversion function can be described in the match statement ,
The local domain and the user searched by the domain search means
If you and The domain is different, you the local domain
Using the conversion function search means for searching the conversion function management table for the conversion function for converting to the domain and the conversion function searched by the conversion function search means, translating the inquiry sentence into an inquiry sentence suitable for each database. However, the gist is to have a search means for searching each database by the translated inquiry text.

According to the first aspect of the present invention, the database storing the search data of the inquiry text input by the user and the local domain corresponding to the search data are searched from the local domain management table, and User domain management from the domain group and user that includes the local domain
To search from the management table, the query sentence to the database
If a conversion function can be written in the and the searched local domain and user domain are different,
Search the conversion function management table for the conversion function that converts the local domain to the user domain , and use this conversion function to translate the inquiry text from the user into the inquiry text that matches each database. Since each database is searched, even if the expression format of the search data in each database is different from the expression format of the user himself / herself, it is possible to search freely. Also,
The present invention according to claim 2 has a plurality of databases.
Have the same meaning in different environments, but
Data items with different expressions
Database for resolving and searching for
Yes, the data representation of the data items in each database
The domain that is the expression is used as the local domain for each database.
The same as the local domain management table that is managed for each
Managing each domain that has the same meaning as a domain group
And within each domain group
Select the standard global domain
Management domain group management table and the global
Between the local domain and the local domain or
A conversion function management table that manages conversion functions that perform one-way conversion.
Users and the domain users can see for each domain group.
User domain that manages main as user domain
Management table and the inquiry text entered by the user
Database storing search data and the search
Set the local domain corresponding to the data to the local domain.
Search in the local management table and
The domain group that contains the main and the user
User domain from the user domain management table
Domain search method to search and inquiry to database
If the conversion function cannot be described in the
Local domain and user domain searched by main search means
If different from the main , change the user domain to local domain
First conversion function to convert to in and local domain
A second conversion function for converting a user domain to the user domain
Conversion function search means for searching from the function management table;
Use the first conversion function searched by the conversion function searching means.
And the value included in the inquiry text is the user domain.
To the local domain and check each database.
The second conversion function is used to search and retrieve the retrieved result value.
To convert from local domain to user domain
The gist is to have means. Main claim according to claim 2
As a matter of fact, if the inquiry text entered by the user is
A database storing search data and the search data
The local domain corresponding to the
Search from the management table and the local domain
Domain groups and users that contain
Search the user domain management table for the database
If the conversion function cannot be described in the query
If this is the case, this searched local domain and user
If different from the main, change the user domain to local domain
First conversion function to convert to in and local domain
The second conversion function to convert the user domain to the user domain
Search from the management table and use this first conversion function
The value contained in the inquiry text from the user.
Convert from main to local domain and
Search results and use the second conversion function
Convert from local domain to user domain.
Therefore, the representation format of the search data in each database is
You can freely search even if it is different from your own expression format.
it can.

[0011] According to a third aspect of the invention, claim 1
Alternatively, in the invention described in 2 , the conversion function searching means is
Direct change between the local domain and the user domain
If the conversion function to be converted cannot be searched, the local
The gist is to have means for searching a combination of a plurality of conversion functions that can convert between a domain and the user domain .

[0012] In the present invention according to claim 3, local
Conversion function to directly convert between user domain and user domain
Local domain and user
Combining multiple conversion functions that can convert between mains,
Convert between local domain and user domain
You can

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a database heterogeneity resolution search device according to an embodiment of the present invention. The database heterogeneity resolution search device of this embodiment is
It has the same meaning in an environment where multiple databases exist, but it searches for data items that have different expressions depending on the database by eliminating the difference in the expressions. The local domain of the database is different from the local domain of the database at the time of searching. In this case, a conversion function between domains is called, a query text of a user is translated into a query text suitable for each database by using this conversion function, and a database is searched by the translated query text. The resolution search device is shown in Fig. 1. In addition, a syntax analysis unit 1 that analyzes the syntax of a query sentence 13 input from the user's application program 12, a query processing control unit 2 that controls the overall flow of processing for this query sentence, and a query conversion that transforms the query. Part 3, a conversion function library 4 that is a collection of functions for converting data between drains, an inquiry processing part 11 that includes a data processing part 6 that converts the domain of retrieved data, and a database for accessing a distributed database An access unit 5, a metadata storage unit 7 that stores and manages information such as domains and domain groups, information about tables included in a database, and metadata that performs conversion of metadata input to the metadata storage unit 7. Management department 8
A plurality of database systems each including a database 9 and a database management system 10 for managing the database 9 are connected to the database access unit 5. The database 9 is also called a local database.

The database heterogeneity resolving search apparatus of this embodiment having the above-described configuration has a two-stage process including a preparation phase for preparing a domain definition and the like before executing a search and a search phase for actually performing a search. Carry out.

First, the preparation phase will be described. In this preparation phase, a domain is defined. The concept of domain management in this embodiment will be described with reference to FIG. As a domain that is a method of expressing the data value of a data item in a database, for example, as shown in FIG. 2, a domain for expressing a blood type includes a blood type value, a blood type code, and a blood type detail. There are various expression methods such as the case of. In this way, by defining those that have the same meaning among various expressions as a domain group and setting only one domain as a domain that can be seen by the user, a single view is provided to the user. can do. In addition, in FIG. 2, the date is also given as a domain group, and the year / month / day and the month. Day. Years, etc. are shown for various domains. Further, in FIG. 2, the user uses the blood group code and the year / month / day, which are the user domain.

Regarding the domain, the following four settings are required: local domain setting, domain group setting, conversion function creation, and user domain setting. The user uses the metadata management unit 8 to set the domain. Set information such as. Further, these pieces of information are stored in the metadata storage unit 7. Further, the configuration of the table of each information stored in the metadata storage unit 7 is shown in FIGS. 3 and 4. The setting of each domain will be described.

(1) Setting of local domain: For the items of the local database 9, a "domain" which is a concept representing the expression format thereof is set. This is called the local domain. For example, for the data item “name” (type is char (20)) of a database, the domain of this is Name
To be set. The meaning of the name domain is not just the type char (20), but the more detailed form of "name". For example, this Name is: -Last name first, first name later.・ The space between the family name and given name is a space.

The expression format is as follows. The definition of the domain for the column is held in the "column" table as shown in FIG.

Basically, this work is performed by the administrator of the local database. At that time, the metadata management unit 8 is used to refer to the already defined domain. If a domain that can be used as a data item domain in your database is already defined, use it and do not create a new domain.

(2) Setting of domain group: Although the format of the expression is different, a group of domains which actually have the same meaning is called a domain group. Domains in this group are all different expressions with the same meaning. In this device, all the domains included in this domain group appear as one domain to the user.

The administrator of the local database determines which domain group the domain defined by him / herself belongs to, and carries out the processing to enter the domain group. If no suitable domain group is found, create a new domain group. If the domain is already defined, you do not need to set it because it is included in any domain group. FIG. 3B shows a domain group.

In the domain group, one standard domain that is used as a standard is set (this setting is basically performed by the administrator of this apparatus). That domain is called the global domain. FIG. 3C shows the global domain.

(3) Conversion function creation: This is not necessary when the conversion function is already defined. A conversion function between the global domain and the local domain is created by using a general programming language. The created function should be in a format that can be compiled and dynamically linked.

The metadata management unit 8 is used to register the relationship between the conversion function and the domain. The correspondence between the conversion function and the domain and the location of the conversion function are held in the "conversion function" table as shown in FIG.

The conversion function creates a bidirectional function such as global domain → local domain and local domain → global domain. However, depending on the conversion, it may not be possible to create a function in both directions (example: name conversion, conversion of "Gengo Suzuki" → "Gengo Suzuki" = blank spaces can be deleted, but the reverse is true). Generally impossible). In that case, create a function in only one direction. The reason for creating a bidirectional function is to combine the functions and perform the conversion when there is no function for the direct conversion.

(4) Setting of user domain: A domain that the user actually sees is selected (for each user) from the domain group. This is called a user domain.
When the user makes an inquiry to the device, even if there are items expressed by various domains, they are all unified by the expression by this user domain. Also, by changing this user domain, the appearance of the data to the user can be changed. FIG. 3D shows the user domain.

Next, the search phase will be described. Search using the domain information as follows. A user generally calls and uses the function of this apparatus in the form of a library from an application program. Inquiries from users are created in the SQL language, which is the search language for relational databases.

(1) Determination of inquiry DB: The syntax analysis unit 1 analyzes the inquiry and finds the database including the table / column included in the inquiry by searching the "column" table of the metadata storage unit 7.

(2) Search for domain of column included in inquiry: FIG. 3 shows domain of column included in inquiry.
Know by searching the "column" table shown in (a).

(3) User domain search: By searching the "domain group user" table shown in FIG. 3D from the inquiring user and domain,
Search for user domain.

(4) Search for conversion function: The conversion function between the user domain and the domain of (2) is checked using the "conversion function" table shown in FIG. When a conversion function that can be directly applied to A → B is not found, it is searched for in combination with A → C and C → B. This is a search problem on a graph in which the domain is a node and the existence of a conversion function between domains is an edge, and can be easily realized by using an algorithm studied in graph theory.

(5) Translation and execution of inquiry: There are two implementation methods depending on the system used for the database access unit 5. The database access unit 5 uses middleware for database integration.

The realization method 1 is a case where the database access unit 5 permits a function like an expandable database.

A database management system whose functions can be expanded by using a function defined by the user is called an expandable database management system (extensible DBMS).
If a function equivalent to this can be used in the database access unit, the conversion function can be realized as a user-defined function of the extensible DBMS. The function can then be used in the query language.

In order to make a query to be passed to the database access unit, the column included in the user's query text is replaced with a form including a conversion function.

[0037] Format. If the domain of the column to be searched and the column of the conditional expression are different from the user domain, perform the following replacement.

A) Replace (search column) with (local domain → user domain conversion function) ((search column)). B) Replace (column included in conditional expression) with (conversion function from local domain to user domain) ((column included in conditional expression)).

The realization method 2 is a case where the database access unit 5 has only the function of the relational database. With middleware that can access the database only with the same function as the relational database, user-defined functions cannot be used in the query language like the extensible database, so by rewriting the query as in implementation method 1, conversion is realized. Can not. It is necessary to realize data conversion as follows.

Conversion of the value contained in the inquiry: The value contained in the inquiry is converted from the user domain to the local domain. Conversion of search result: The data processing unit 6 converts the value of the search result from the local domain to the user domain.

In each method, a plurality of databases are accessed, and the result is the sum of search results for each database.

Next, two examples of the case where the synthetic conversion is performed and the case where the synthetic conversion is not performed will be described. FIG. 5 shows a sample database (DB) used. For metadata such as domains, the values in Figures 3 and 4 are in line with this example. The same database is used in both Example 1 and Example 2. The two examples are common until the definition of the conversion function in the preparation phase. The destination is different from the user domain setting. The two databases (DB) are in the following formats.

[0043] That is, the "name" of DB1 does not leave a space between the surname and the name, and the "name" of DB2 puts a half-width space between the surname and the name. The blood type of DB1 is represented by a code, DB2
The blood type of is represented by a value.

First, as Example 1, the case where the synthesis conversion is not performed will be described. As the preparation phase of Example 1, the local domain, the domain group, the definition of the conversion function, and the determination of the user domain are performed as described above, respectively, as follows.

(1) Definition of local domain: DB1,
The DB2 administrator sets the local domain for each DB column. Since there is no space between first and last name in the domain of full name in DB1, it is NameNoSpace. Since the blood type domain of DB1 is represented by a code, it is BloodTypeCode. Since there is a space between the surname and surname in the domain of the name of DB2, it is named NameSpace.
Since the blood type domain of DB2 is expressed by a half-width value, it is BloodTypeValue.

(2) Definition of domain group: Here, as a domain group, ・ Name (the included domain is NameNoSpace, NameSpac
e, global domain is NameNoSpace), BloodType (included domain is BloodTypeCode,
BloodTypeValue, BloodTypeValueLarge, and global domain are BloodTypeValueLarge). Here, BloodTypeValueLarge is a domain having a value in double-byte characters (A, B, O, AB).

[0047] (4) Determination of user domain: In this example, the user domain is equal to the global domain (FIG. 3 (d)).
This is the case when the user gsuzuki in. That is, And

Next, the search phase of Example 1 will be described. ◎ Convert user inquiries into SQL sentences as follows. Inquiry: select name where blood type = “A” (1) Search for table containing blood type and name: Search “column” table to find DB1 and DB2.

[0049] In the following (4) and subsequent steps, the above-described realization method 1 and realization method 2 will be described separately.

The case of implementation method 1 will be described. (6) Query execution: The search results for each DB are as follows. DB1: Miki Suzuki DB2: Hiroki Machihara The final search result is the sum of the results, which is Miki Suzuki Hiroki Machihara.

Next, the case of the realization method 2 will be described. (6) Translate the inquiry.

DB1: select name where blood type = "0" DB2: select name where blood type = "A" Becomes

Next, as Example 2, description will be made on the case where the synthetic conversion is performed. Consider a situation where a user who uses DB1 wants to obtain information in the format of DB1. At this time, the user domain is set to the DB1 type (this is the case when the user satoh in FIG. 3 uses it).

First, the preparation phase will be described.
The procedure up to (3) is the same as in Example 1, so (4) is performed as follows.

(4) Determination of user domain: The user domain of Name is NameNoSpace. Let BloodTypeCode be the user domain of BloodType. Next, the search phase of Example 2 will be described. The user's inquiry is changed to the SQL statement as follows.

As described above, according to the present invention, the database storing the search data of the inquiry text input by the user and the local domain corresponding to the search data are searched from the local domain management table, and A user domain is created from the domain group and users that include the local domain.
Search from the management table and query the database
When a conversion function can be written in a statement and the searched local domain and the user domain are different, a conversion function for converting the local domain to the user domain is searched from the conversion function management table, and this conversion function is searched. Is used to translate the query text from the user into a query text that is suitable for each database, and the translated query text is used to search each database, so the expression format of the search data in each database is the same as the user's own expression format. Even if it is different, the user can freely search and easily obtain the data in the desired expression format, and can use the data in the database more freely. In addition, it is possible to eliminate the difference in expression format without the need to define the global schema. In addition, in the present invention
According to the search data of the inquiry text entered by the user.
Data stored in the database and the search data
Replace the corresponding local domain with the local domain management table.
Search from Bull and include the local domain
User domain from the domain group and user
Search from user domain management table to database
When a conversion function cannot be written in the
So, this searched local domain and user domain
Is different from, change the user domain to the local domain
The first conversion function to convert and the local domain are used.
The second conversion function for converting to the domain
Table, and using this first conversion function,
Is the value included in the inquiry text from the user domain
To a local domain and search each database
Then, the retrieved result value is read using the second conversion function.
Since each domain is converted from the local domain to the user domain,
The expression format of the search data in the database is the user
Even if it is different from the expression format of
You can freely search the data with the formula and get it easily.
Database data can be used more freely
Become. Also, without the need to define a global schema,
It is possible to eliminate the difference in expression format.

In addition, in the existing multi-database, the difference in expression format cannot be resolved without knowing the expression format in all databases, whereas in the present invention, the administrator of the local database has it. Only the local domain, which is the expression format in the database, and the global domain, which is the standard expression format, need to be considered, and it is not necessary to be aware of how they are expressed in other databases.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a database heterogeneity resolution search device according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing a concept of domain management.

3 is a column table stored in a metadata storage unit used in the database heterogeneity resolution search device of FIG. 1;
It is a figure which shows the structure of a domain table, a domain group table, and a domain group user table.

FIG. 4 is a diagram showing a configuration of a conversion function table stored in a metadata storage unit used in the database heterogeneity resolution search device of FIG. 1.

FIG. 5 is a diagram showing a sample database.

FIG. 6 is a diagram showing an example of a blood type correspondence table used by a conversion function.

FIG. 7 is an explanatory diagram showing a synthesis conversion of domains.

[Explanation of symbols]

1 Parsing part 2 Inquiry processing control unit 3 Inquiry converter 4 Conversion function library 5 Database access department 6 Data processing unit 7 Metadata storage section 8 Metadata management section 9 Database 10 database management system

Continuation of the front page (56) Reference JP-A-7-65032 (JP, A) JP-A-6-67867 (JP, A) Nishizawa Satoshi, Takasu Atsuhiro, Adachi Atsushi, schema integration and query processing in information retrieval, Information Processing Society of Japan Research Report (93-DBS-9-4), Japan, Information Processing Society of Japan, July 23, 1993, Vol. 93, No. 65, p. 69-77 DURAND, J. et al. , Object View Broker: A Mediation Service and Architecture to Provided Object-Oriented Views of, Computer Science, July 1994, Vol. 819, p. 412-428, Continuation of title: Heterogeneous Databases (58) Fields investigated (Int.Cl. ⁷ , DB name) G06F 17/30 G06F 12/00 JISST file (JOIS)

Claims (3)

(57) [Claims] 1. A database heterogeneity resolving search device for retrieving data items having the same meaning in an environment having a plurality of databases, but having different expressions depending on the databases, by resolving the differences in the expressions, and each database comprising: The local domain management table that manages each database as a local domain, which is the data representation format of the data items of, and each domain that has the same meaning is managed as a domain group, and the standard in each domain group is standard. Between the global domain and the local domain, and a domain group management table for selecting and managing the globally used global domain.
Yu and conversion function management table for managing the transformation functions for bidirectional or unidirectional conversion, the domain where the user looks for each domain group
User domain management table to manage as user domain
And a database storing search data of inquiry texts entered by a user and a local domain corresponding to the search data from the local domain management table, and a domain group including the local domain and the user. From user domain before
The domain search means to search from the user domain management table , and the conversion function can be described in the inquiry text to the database.
A case that can, if the local domain and a user domain searched by the domain search means are different, the conversion function search means for searching for a conversion function that converts the local domain to the user domain from the conversion function management table, the And a search unit for translating the inquiry sentence into an inquiry sentence suitable for each database using the conversion function searched by the conversion function searching unit and searching each database by the translated inquiry sentence. Database heterogeneity resolving search device. 2. In an environment having a plurality of databases
Have the same meaning, but depending on the database
For data items with different expressions, eliminate the difference in the expressions.
A database heterogeneity resolving search device for searching, which is a data representation format of data items in each database.
Domain as local domain for each database
Each domain that has the same meaning as the local domain management table to be managed as a domain group
Manage and represent within each domain group
And select the standard global domain
Between the domain group management table to manage and the global domain and the local domain
A conversion function that manages the conversion function that performs bidirectional or unidirectional conversion of
The conversion function management table and the domains that users can see for each domain group
User domain management table to manage as user domain
And the search data of the inquiry text entered by the user are stored.
Corresponding to the database and the search data
Local domain to the local domain management table
Search from and include the local domain
User domain from domain group and said user
Domain search to search from the user domain management table
A conversion function can be described in the search means and in the inquiry statement to the database.
If the domain search method is not available,
If the local domain and the user domain are different,
The first conversion function for converting the domain to the local domain
Number, and convert local domain to user domain
The second conversion function is searched from the conversion function management table.
And a conversion function searching means for searching for the first conversion function searched by the conversion function searching means.
Use the value included in the inquiry text to
Database to local domain
Search and use the value of the retrieved result by the second conversion function
To convert from local domain to user domain.
Database heterogeneity characterized by having search means
Resolution search device. 3. The conversion function searching means is the local
A conversion function that directly converts between the domain and the user domain.
If the number cannot be searched, then the local domain and the previous
3. The database heterogeneity resolving search device according to claim 1, further comprising means for searching a combination of a plurality of conversion functions capable of converting between user domains .