JPS62269244A - File access system - Google Patents

Abstract

PURPOSE:To increase an access speed, and also, to reduce the capacity of a memory with respect to the same record, by constituting a data having a linkage logically, of an index use file and plural data files, and executing an access to plural data files by using the index use file. CONSTITUTION:With regard to (m) pieces of data A1-Am provided with the same attribute, data groups having a linkage logically, respectively are arranged to (a11-a1n), ..., (am1-amn) in accordance with an access purpose, and a file F1 consisting of (m) pieces of records a11-am1 is generated for each access purpose. In the same way, a file F2 consisting of (m) pieces of records a12-am2 is generated, and thereafter, successively, a file Fn consisting of (m) pieces of records a1n-amn is generated. On the other hand, in order to maintain a logical linkage of (m) pieces of records in F1-Fn and the data A1-Am, an index use file FX in which a relative pointer corresponding to (m) pieces of records, namely, corresponding to (a11-a1n), ..., (am1-amn) is provided. By an access of the index use file FX, a data file is brought to an access.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a file access method in a computer system in which a large number of file accesses occur to the same file and the memory capacity is limited.

(Prior Art) An example of a conventional file structure in a computer system is shown in FIG.

With reference to FIG. 4, m pieces of data A having the same attribute
m data groups (a11 to a1n) to (am1 to amn) that are logically connected with respect to 1 to Am, respectively.
One file F is composed of each code.

Here, the above-mentioned file access will be explained. Referring to FIG. 5, access program P11 accesses data A.
Access data group all related to 1 (shaded area)
When accessing the data A1, the access program P11 must access all the data groups related to the data A1, that is, the entire code must be accessed.

The access buffer memory BF for the access program P,1 requires a capacity for the code of the file F, that is, all the data portions related to the data A1.

Furthermore, when the access programs P, 1 to Pan simultaneously access the data groups a11 to a1n related to the data A1, each requires an access buffer memory similar to that described above. Furthermore, when the access programs P, 1 to P1n access simultaneously, the accesses are concentrated on one record (the record related to data A1), resulting in serial processing due to record control.

(Problems to be Solved by the Invention) As mentioned above, in the case of the conventional file access method, since the entire record in which access target field 1 exists according to the access purpose is accessed, the record of file F The larger the length and number of accesses, the larger the capacity of the access buffer memory and the greater the number of access buffer memories required. Also, if the number of accesses is large, there will be waiting time due to record exclusive control.

There is a problem that the processing time becomes long.

A file access method according to the present invention is based on a file access method in which a file device stores a plurality of logically connected data groups with the same attribute regarding predetermined data according to the access purpose. The present invention is characterized in that it has a data file and an index file in which a record relative pointer to the data file is stored, and the data file is accessed by accessing the index file.

(Example) The present invention will be explained below with reference to Examples.

First, the file structure used in the present invention will be explained with reference to FIG.

Regarding m pieces of data A1 to Am having the same attribute, logically connected data groups are respectively (a11 to a1n), . . . , (am1 to
amn) to create a file F1 consisting of m X codes a11 to am1 for each access purpose.

Similarly, a file F2 (not shown) consisting of 2m records a12 to am2 is created, and thereafter files up to a file Fn consisting of 1m records a1n-amn are created in sequence.

On the other hand, m records in Fl~Fn and data A1~
for m records to maintain logical connection with Am. That is, (a11-a1n),...

An index file FX is provided in which relative pointers to (am1 to amn) are stored.

Next, file access will be explained with reference to FIGS. 2 and 3.

As shown in FIG. 2, when the program P1n accesses the data group a1n related to the data A1, the program Pan first accesses the index file FX. Then, a pointer corresponding to the data A1 (that is, to the data group (a11 to a1n) related to the data A1) is obtained, and based on this pointer, the corresponding data group a1 of the target file Fn is selected from the access purpose-specific files F1 to Fn.
n is accessed, and this data group a1n is stored in the buffer memory BF1.

On the other hand, as shown in FIG.
Each Pan is a data group a11~ related to data A1.
When accessing a1n simultaneously, programs P11 to P
an accesses the index file FX.

get pointer. Program P,1 accesses data group a11 based on this pointer, and
accesses the data group a12 based on this pointer. Similarly, the program Pan accesses the data group a4n based on the pointer.

In other words, the f programs P11 to P1n access the files F1 to Fn, respectively, and can be accessed without being subjected to record exclusive control.

It should be noted that all accesses to the file FX by the programs P11 to P1n are made to the same record, but these accesses are for reference and are therefore unrelated to record exclusive control. Further, the capacity of the access buffer memory BF required for the programs P11 to P1n is equal to the data length of a single group when accessed by each program.

Furthermore, the capacity of the buffer memory used for accessing the file FX is small (effect of the invention).As explained above, in the present invention, logically connected data is composed of an index file and multiple data files, and the index file Since multiple data files are accessed using , the access speed for the same record is fast and the memory capacity is small.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the structure of a file according to the present invention. Figures 2 and 6 are diagrams for explaining access according to the present invention, Figure 4 is a diagram showing the conventional file structure, and Figure 5 is a diagram showing the conventional file structure.
The figure is a diagram for explaining conventional access. A1-Am... m pieces of data with the same attribute, a1i
~a1n...A data group of logically connected data groups for A1 classified according to n access purposes+
am'~amn...A data group in which logically connected data groups regarding Am are classified according to n access purposes, F1~Fn...Files classified by access purpose,
FX...File for index, P11~Pan...a+
A program that accesses each of i to a11, B
F, ~BFn...P+1~Buffer memory for access in Pan, F...File. Figure 1 Fl Figure 2 Figure 4

Claims (1)

[Claims] 1. In a computer system equipped with a file device, the file device stores a plurality of data files in which groups of logically connected data with the same attributes regarding predetermined data are stored according to access purposes, and records for the data files. 1. A file access method comprising: an index file in which a relative pointer is stored; and the data file is accessed by accessing the index file.