JPS62288934A - Massive information retrieving system - Google Patents

Abstract

PURPOSE:To collectively store values of a designated column by designating the column of a data base retrieval key, which is effective for retrieval, to a control system at the time of data base constitution to determine the storage position on the data base of each information. CONSTITUTION:A console 1 for system manager and plural user terminals 3 which users operate are connected to a data base control system 4 of a mass information retrieving system. A key information ID control system table 7 is provided in the system table group of the system 4, and an external data base 5 is connected to a buffer 6. A manager 2 of this retrieving system designates the column of the data base retrieval key, which is effective for retrieval, to the system at the time of constituting the data base 5. The position on the data base 5 of each information is determined, and values of the designated column are collectively stored in the buffer to retrieve plural pages.

Description

[Detailed Description of the Invention] 3. Detailed Description of the Invention [Field of Industrial Application] The present invention provides a system for searching a large amount of information using a database, in which the information handled has large categories, and This invention relates to a relational database management method suitable for cases where there are keys that characterize a database.

[Conventional technology]

The conventional device is a CP that performs view processing when applied to views, as described in Japanese Patent Application Laid-Open No. 56-21273.
In order to reduce U time and the number of file accesses, methods such as creating view-compatible data have been used.

However, no consideration has been given to reducing the number of accesses to the relational database by taking into consideration the data storage location on the secondary storage device.

[Problem that the invention seeks to solve]

In the above conventional technology, in a relational database management system, "the location on the secondary storage device where data is stored is determined by the database management system."J
Information is processed and stored according to the principle of
Due to the relationship between the buffer size of the database management system that takes in information in the database and the size of the database, in systems where the latter is very large, the number of accesses to the database increases, leading to performance deterioration. there were.

An object of the present invention is to reduce the number of accesses to a database and prevent performance deterioration when searching for information on a database containing a large amount of information.

[Means for solving problems]

``In a relational database management system in an information retrieval system, when building the database and when adding or changing information, the location where the information is stored is automatically determined in the same location for similar information, This is achieved by storing. In other words, the information retrieval system administrator has designed the database to be effective for a larger number of searches based on the main search patterns of the information retrieval system users and the nature of the information stored in the database. By specifying a search key field to the database management system when constructing the database, the two-part relational database management system can store the information on the database when creating the database and adding, deleting, or modifying the information. This is achieved by storing the storage locations for each value in the specified field and creating a system table for managing the storage status of the information.

[Effect]

The database management system of the above information retrieval system is
''Regarding the value of the key column of the information storage table designated by the administrator of the two-part information retrieval system, records are collectively stored for each column value of the information registered in the database. In other words, the database management system has one buffer surface (
The size of a page (hereinafter referred to as a page) is taken as one unit, and management is performed so that only information having the same column value exists within the same page. When adding to or updating the database, an information ID that is a pair of a page number and the address of the page is registered in the key-based information ID management system table as the storage location of the information.

Furthermore, the information retrieval system administrator specifies several secondary keys that are often used for information retrieval of the database and have a lower priority than the above key.
A reverse file list is created on the system table, in which IDs of information having the same secondary key field value are connected by pointers. When updating the data page-]2 information, if the field value of the above primary key is updated, the above key information ID
Refer to the management system table, re-storage to the appropriate location, and correct the information ID on the system table. When the value of the secondary key is updated, the reverse file list is corrected by changing the pointer. With this method, when searching for information, the database management system can know the page that stores the information requested by the user of the information search system by referring to the above system table, and can search for the minimum necessary information. In the present invention, the greater the duplication rate of the specified key value, the greater the effect.

When a page overflows, the page closest to the page is set as the next data storage page.

According to the present invention, since an index can be created on a page-by-page basis, it is possible to reduce the capacity of the index.

〔Example〕

An embodiment of the present invention will be described below with reference to FIG. That is, the information retrieval system includes a system administrator 1 who manages the information retrieval system, and a console such as a CRT display 2 that mediates communication between the system administrator 1 and the database management system 4 in order to operate the information retrieval system. , a terminal that mediates communication between the information retrieval system user and the database management system 4, such as a CRT display 3, and a database that stores information searched by the system user, that is, 2.
A storage device 5, a buffer 6 for receiving and receiving information accessed from the database in the database management system 4 that manages the database, and a key-based information ID management system table 7 created by the database management system.
Consists of etc. ``Two-Part Information Retrieval System Administrator'' becomes the primary key and secondary key when creating the database, that is, when designing the table. The column names of the tables in the database are communicated to the database management system. In this embodiment, it is assumed that information as shown in FIG. 2 is stored. That is, the document information includes document number 5-1, document classification 5-2, document name 5-3, author name 5-4, author's affiliation 5-5, publisher name 5-6, document publication date. 5-7, Document content keyword 5-8, Document content summary 5-9, Document purchase date 5-10, Document purchase amount 5-11. The table has columns for document storage location 5-12 and document management information 5-13. The system administrator 1 first selects the document number column 5-1 to the document content summary column 5 because the column value information after the document purchase date 115-1-0 in the table is not disclosed to the user as document management information. -Define a view for 9. Next, due to the nature of the information, by selecting the document classification column 5-2 as the primary key, the document information will be stored in the database 2 for each category, and at the same time the document name and author name. , information in which most of the content keywords of documents have the same value will be stored within the same page. As the secondary key, specify the publisher name field 5-6 that cannot be classified using the two-legged primary key, and the remaining 5-3 to 5-5 and 5-7.
．． Indexes are provided for each column 5-8. By designing the table as described above and transmitting it to the database management system 4 at the time of database construction, the database management system 4 can create the key-by-key information ID management table 7 and the primary key information ID management table 7 regarding the primary key shown in FIG. 3 when inputting the document information. 2
Create a reverse file list for the next key. ” When a user of the two-part search system inputs information search conditions from the terminal 3, the database management system 4 extracts the information requested by the system user using the above system table and index. Provide to users. When updating or deleting information,
This is performed by the system administrator 1, and the database management system 4 modifies the key-based information ID management system table each time. With this method, the information requested by users of the two-part information retrieval system exists on a page-by-page basis, reducing the number of database accesses and improving search performance.
Hi.

〔Effect of the invention〕

According to the present invention, the information retrieval performance is improved in a large-volume information retrieval system having a relational database. The information retrieval efficiency of the database depends on the selection of the primary key and the nature of the information, but generally speaking, the larger the amount of information, the higher the information retrieval efficiency. For example, if one piece of information has an average length of n bytes, and N pieces of information are stored in a database, and the buffer length (size of one page) is B bytes, one page contains [Nn/Bl -1 pieces of information. can be stored. In a conventional database, the average percentage of information with different primary key values stored per page is M
%, there will be information with the same key value across (1-00-M) /N pages, and at least (1
00-M)/N pages needed to be accessed. In the method of the present invention, this can be accomplished by searching [(100-M)n/Bl pages. Here, the symbol [a] represents the smallest integer greater than or equal to d.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram illustrating an information retrieval system according to an embodiment of the present invention, FIG. 2 is an explanatory diagram of a literature table as an example of a table that stores information handled by the information retrieval system, and FIG.
FIG. 1 is an explanatory diagram of a key-based information ID management system table that improves the effects of the present invention. 1... Information retrieval system administrator, 4... Database management system, 5... Database, 6... Buffer, 7... Information by key JT) Management system table,
5-1 to 5-13・Reference table column. 7de°1--8 (Pa

Claims (1)

[Claims] 1. In a table representing information in a relational database, the administrator of the information retrieval system specifies a key column, and the system uses this column to create a database and to add, delete, or change information on the database. 2. Configure the database for each field value above.
The storage location on the secondary storage device is stored in the same location in units of accessing information from the secondary storage device, and the storage location is automatically created and modified in memory as one of the system tables, and the information A mass information retrieval system comprising a relational database management system having a function of accessing a database by using the above table when searching for.