JPH04323743A - Null key excepting index management processing system - Google Patents

Abstract

PURPOSE:To efficiently retrieve a data base in the index system by suppressing index registration to invalid data for a user. CONSTITUTION:A null value invalidating attribute 2 is given as the key attribute for generation of an index key 1 and is set as the attribute (property) of the index part after generation. When the index part 1 having the null value invalidating attribute 2 will be generated, all records 7 of a data file 6 are read in, and values of key fields 8 of records 7 are evaluated in order from the smallest value, and a key entry 4 of the key field 8 whose value is not null is registered in an index block 5 but that of the key field 8 whose value is null is not registered. When a record will be added to the already generated index part 1, it is discriminated whether the value of the key field of the record is null or not, and it is not registered in the index part 1 if it is null.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] This invention provides an index that reduces index search time by not registering redundant key information in the index in a database processing system that has an index file that associates key values with record numbers. This invention relates to a management processing method.

0002

2. Description of the Related Art FIG. 2 shows a method for accessing a data file at high speed, for example, by using a specific item in a record described in a document (Mitsubishi Office Computer MELCOM80DPS10 Data Management Manual) as a key, and linking that key with the record. FIG. 4 is a diagram illustrating a method of creating a correspondence table as an index and using this index to access records corresponding to arbitrary key values at high speed. In the figure, 1 is the index part of the index file, 3 is the root index block in the index part, 4 is the key content in the index block, 5 is the index block placed in the lowest layer, 6 is the data file, and 7 is in the data file record, 8 is a key field within the record.

Next, the operation will be explained. To generate the index part 1, read all the records in the data file 6, rearrange the values in the key field 8 in descending order, and write the key value and record position information (record number) in the index block 5 as the key content 4. register. In the data search process, in order to search for a key entry 4 having a specified key value, a lower index block 5 containing the specified key value is obtained from the root index block 3, and a lower index block 5 containing the specified key value is obtained from the index block 5. The key record 4 can be searched and the corresponding record 7 can be searched from the key record 4.

0004

[Problems to be Solved by the Invention] In the conventional system, key entries corresponding to all records are registered in the index section, which causes the following problems. (b) If you generate an index that includes data that is meaningless to the user in the key field (for example, data that will be left blank for the time being to be used for future expansion), the same value (blank data) will be created. A large number of key entries are created, and it takes time to search the database. (b) Similarly, adding and deleting records takes time.

The present invention has been made to solve the above-mentioned problems, and has the following objects. (b) A null value invalid attribute is provided as an attribute of the key field, and when creating an index or adding a record, if the data corresponding to the key field has a null value, it will not be registered in the index and unnecessary key registration will be performed. to enable high-speed index searches. (b) By reducing unnecessary key registrations, it is no longer necessary to delete keys with null values in record deletion operations, and the performance of deletion operations can also be improved.

[0006]

[Means for Solving the Problems] The index management method according to the present invention specifies whether or not to make a null value invalid in the attribute definition of a key field when generating an index. In this case, the information is recorded as index management information.

[0007] Also, null values include blanks if the key field is a character type, and NULL values if the key field is a numeric (excluding binary) type, and the key value is evaluated during index generation and null keys are excluded. It is something to do.

[0008]

[Operation] In the database search using the index method according to the present invention, by managing only the keys of data records that are meaningful to the user, it is possible to reduce redundant index searches and perform database searches at high speed.

[0009]

【Example】

Example 1. Hereinafter, a diagram of an embodiment of the present invention will be described. In FIG. 1, 1 is the index part of the index file, 2 is a null invalid attribute in the index part, 3 is the root index block in the index part, 4 is a key content in the index block, 5
is an index block placed in the lowest layer, 6 is a data file, 7 is a record in the data file, and 8 is a key field in the record.

[0010] The null value invalid attribute 2 is given as a key attribute when generating the index section 1, and after generation, it is set as an attribute (property) of the index section.

To generate the index section 1 having the null value attribute 2, all records 7 of the data file 6 are read, and the values of the key fields 8 of the records 7 are evaluated in descending order.If the value is not null, the index block 5 is Navel key control 4 is registered, and if it is a blank value, key control 4 is not registered.

[0012] Also, in the operation of adding record 7 to the index section 1 that already exists, the key field of that record is evaluated to see if it is a null value, and if it is a null value, it is not registered in the index section. do not have.

[0013]

As described above, according to the present invention, by suppressing the index registration of data that is invalid for the user, the number of registered key entries in the index section can be reduced to only the necessary ones, and an efficient database using the index method can be created. This has the effect of making search possible.

[Brief explanation of drawings]

FIG. 1 is a diagram showing an index management processing method according to an embodiment of the present invention.

FIG. 2 is a diagram showing a conventional index management processing method.

[Explanation of symbols]

1 Index section 2 Null value invalid attribute 3 Root index block 4 Key control 5 Index block 6. Data file 7 Record 8 Key field

Claims (1)

[Claims] 1. Index management in a system that manages a database consisting of a data file that holds actual data and one or more index files that store index information using specific items in the actual data as keys. In this method, the keys in the index file have a one-to-one relationship with each record in the data file.
If the corresponding key field in a record is meaningless data for the user (null data), it has a means to suppress the registration of the key corresponding to that record in the index file. A null key exclusion index management processing method characterized by: