JPH03147150A - Data processing system - Google Patents

Abstract

PURPOSE:To shorten time for a data processing by rearranging input records in the order of values provided for items in advance by a central processing unit and executing an output record preparation processing while skipping a reference instruction when a previously read key value is coincident with a key value to be read this time. CONSTITUTION:A central processing unit 10 reads an inputted file intermediate file 42 and moves a field A of an input file 41 to a new key area 21. Next, the new key area 21 is compared with an old key area 22 and when they are same, the file is already read the previous time or the time before the previous time. Therefore, the next output record preparation processing is executed without referring a reference file 30. Thus, the number of times for accessing the reference file with the data processing, namely, access time is reduced and the time for the data processing of a program is shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a data processing method, and more particularly to a data processing method in which other files are referenced by predetermined items in an input record.

[Conventional technology]

Conventionally, in this type of data processing method, when referring to another file based on a predetermined item in an input record, the other file is referred to unconditionally without any particular consideration.

Therefore, the same number of other files as the number of input records were referenced.

[Problem to be solved by the invention]

In the conventional data processing method described above, since other files are referenced unconditionally, the data processing time is directly dependent on the number of input records, and the disadvantage is that the data processing time becomes longer when the number of input records is large. There is.

[Means to solve the problem]

The present invention provides a data processing method for a computer system that refers to other files based on a specific item in an input record, in which a central processing unit sorts the input record in advance into one item number of the value of the item, Immediately before the instruction to reference the other file, check whether the key value read previously matches the key value to be read this time,
If they match, the reference command is skipped and output record creation processing is executed.

〔Example〕

Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of hardware for realizing the data processing method of the present invention. In FIG. 1, 10 is a central processing unit, and 20 is a main memory, in which a new key area 21 and an old key area 22 are secured. 30 is a reference file, which exists on the magnetic disk device. 41
・42 and 43 are input files, intermediate files, and output files, respectively, and magnetic tape devices are used.

FIG. 2 is a diagram illustrating the job flow of the hardware shown in FIG. 1. In FIG. 2, the explanation of each stem knob is as follows.

5TEP I: By sorting (classifying) the input file 41 in the order of the values in the field A/area using "Field A" as the sort key, input records with the same value in the field A/area are successively stored in the intermediate file. 42. This improves the hit rate of the key check routine in the next step. Note that since step I can be executed using a service program, it is possible to use the job control language (JCL). Only additional coding is required.

5TEP II: This is a data processing program created by the user, which inputs the intermediate file 42 sorted by "field A" of the input file 41, refers to the reference file 30 using the value of field A/area, and generates an output file. This is a program that creates an output record to be output to 43.

FIG. 3 is a flowchart showing the data processing performed in FIG. In FIG. 3, the flow surrounded by broken lines is the part that has been changed from the conventional flow by adopting the present invention. First, the central processing unit 10 reads an input file (intermediate file 42 in FIG. (102). Next, compare the new key area 21 and the old key area 22 (this is the area name of the area that stores the key value when the previous reference file 30 was referenced). +03> If they are the same, If there is, it has already been read the previous time (or the time before the previous time), so the next output record creation process is executed without referring to the reference file 30 (1θ6). Also, if the new key area 21 and old key area 22 are not the same, read the reference file 30 (104>, change the day key value to the new key value (+05>), and then start the next output record creation process. Execute (+06). Note that (old key value) ≠ (
Since the number of times the new key value) is already performed in step I of Fig. 2, the maximum number of times the reference file
is less than or equal to the number of keys (number of data) it has. That is, the reference file is referenced only once for one key value.

〔Effect of the invention〕

As explained above, according to the data processing method of the present invention,
Previously, the reference file was referenced the same number of times as the number of input records (for example, m), but the maximum number of keys that the reference file has (same as the number of data, for example, n
) The following is true. That is, the reference file is referenced only once for one key item. For this reason, m
In the case of an, the greater the number of input data compared to the number of reference data, the greater its effectiveness.
Even in this case, if there is data with the same reference key in the input data, the number of times the reference file is referenced is reduced compared to the conventional method.

Therefore, the number of times the reference file is accessed during data processing, that is, the access time is reduced, and the data processing time of the program can be reduced.

- As an example, the number of input records is 300,000, and the number of reference file data is 2,000 (m=300,000, n=2.00
0>, the process that took about 8 hours in the conventional method can be reduced to 1/8, or about 1 hour, including sorting.

[Brief explanation of the drawing]

Figure 1 is a block diagram showing an embodiment of hardware for realizing the data processing method of the present invention, Figure 2 is a diagram explaining the job flow of the hardware in Figure 1, and Figure 3 is the diagram in Figure 2. 3 is a flowchart showing data processing executed in the process. 10... Central processing unit, 20... Main storage device, 21
...New key area, 22...Old key area, 3
0...Reference file, 41...Input file, 42
...Intermediate file, 43...Output file.

Claims (1)

[Claims] In a data processing method for a computer system that refers to other files based on specific items in an input record, the central processing unit sorts the input records in advance in the order of the values of the items, and Immediately before a reference command, it is checked whether the key value read previously matches the key value to be read this time, and if they match, the reference command is skipped and output record creation processing is executed. data processing method.