JPS593549A - File management system - Google Patents

Abstract

PURPOSE:To shorten a retrieving time of an index file, by placing node of continuous level at constant interval or more, when assigning each node of a wooden structure to an auxiliary storage device. CONSTITUTION:A node 8 of the second level is placed so that an inerval to a node 7 of the root exceeds the time required for retrieving a data in the node of the root at the processing device side. The interval between a node 9 of the third level and the node 8 of the second level, and its subsequent one are placed in the same way of thinking. In this way, when an index file is referred to, extending from the node of the root to a node of a leaf, waiting for access rotation by an auxiliary storage device is not caused in the process shifting to processing of a node of the following level, and a retrieving time of the index file can be shortened.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a computer system, and particularly to a file management method suitable for searching data on an auxiliary storage device.

When a computer system handles a large amount of data, this data is stored on a large-capacity auxiliary storage device. The data transfer speed between the auxiliary storage device and the main storage device is slower than the speed of the processing device, and a file management function is required to efficiently write and read data.

FIG. 1 shows an example of a file structure created in an auxiliary storage device. A file 1 is composed of a collection of records 2, which are units for handling information. When searching for an arbitrary record 2 in file 1, if a method is adopted in which the records are sequentially read and checked from the first record 2 to the last record 2, it becomes necessary to search half of the total number of records on average. For this reason, an index file 3 that stores index-only information is provided so that necessary records 2 can be searched for more quickly.

FIG. 2 shows the logical structure of the index file 3. The index file 3 is created using a data structure method having a tree structure. The tree structure consists of node 4 at the branch point and each node 4
It consists of a branch pointer 5 to the next lower level stored in fL. The branch pointer 5 at the lowest pel node (called the leaf node) is empty and the record pointer 6 is used instead.
have.

The search for record 2 is to obtain the corresponding record 2 given information (called a key) that serves as a clue for the search. In this method, this can be achieved by setting the pointer from the highest level node (called the root node) to the leaf node according to the key information to find the record pointer 6. .

FIG. 3 shows the structure of an index file on an auxiliary storage device in the prior art. They are arranged in the order of nodes 7 at the root, nodes 8 at the second level, and nodes 9 at the third level in accordance with the rotational direction of the auxiliary storage device. In this case, the contents of the root node 7 are read to the processing device side, the data of the node circle is determined, a pointer to the next level node 8 is obtained, and a corresponding read instruction is issued to the auxiliary storage device. Until then, the storage position of node 8 is at the read head position of the auxiliary storage device t.
It passes through i1tO. Therefore, reading node 8 requires waiting until the auxiliary storage device rotates once and returns to the head position IO. This relationship is shown in the time chart of FIG. This problem occurs at each level, and has the drawback of increasing the search time for index files.

An object of the present invention is to provide a file management method that speeds up file management processing and improves the processing performance of a computer system by shortening the search time of the Intex 7 file.

As a means of reducing index file search time, the present invention allocates each node of this structure to an auxiliary storage device, and the nodes of subsequent levels require the time required for the processing device to search for node data. By taking this into consideration, they are arranged at a certain interval or more, so that unnecessary waiting for access rotation of the auxiliary storage device does not occur when searching for an index file.

FIG. 4 shows the structure of an index file when the present invention is applicable. The second level node 8 is arranged such that the distance from the root node 7 exceeds the time required for the processing device 110111 to retrieve the data in the root node 7. The spacing between the third-level node 9 and the second-level node 8 and the subsequent nodes are arranged in a similar manner. This results in
When indexing an index file from the root node 7 to the leaf nodes, there is no longer a wait for access rotation to the auxiliary storage device in the process of moving to the next level of nodes, and the index 7 al. This has the effect of shortening search time. This relationship is shown in the time chart of FIG.

FIG. 5 shows a processing flow of an embodiment of the present invention. An auxiliary storage allocation area free management table 11 is secured on the main storage or auxiliary storage as a table for managing the allocation of each node of the index file having a tree structure on the auxiliary storage. Vacant management table 11
is configured to store the empty or used state of this area in correspondence with the rotational position of the auxiliary storage device 120, and if the content of the table corresponding to the area is 00, it is empty, 1
indicates that it is in use. The free space management table 11 is used when constructing a tree structure of index files on the auxiliary storage device 12, and the initial state is all O's. When each node of the tree structure is allocated on the auxiliary storage device 12, the corresponding portion is changed from 0 to 1. The procedure for allocating each node of this structure on the auxiliary storage device 12 is as follows.
The flow is shown in FIG. Layout of the entire tree structure is completed by repeatedly executing the processing of this flow.

According to the steps in this flow, in step 13, the time required for searching the nodal circle data on the processing device side is determined. In step 14, the allocation position on the auxiliary storage device of the node above the node to be allocated this time is determined.

In step 15, the location where the currently allocated node should be stored is calculated from the results of steps 13 and 14. At step 16, it is checked with reference to the table 11 whether or not there is a build-up at the corresponding position.

If there is no free space, the table reference pointer is updated in step 17 so that the planned position is advanced to the next rotational position. If a vacancy is found at 16,
At step 18, a message indicating that the relevant position is being managed is displayed in the table 11, and at step 119, the section information is stored at the relevant position in the auxiliary storage device. This completes the assignment to one node.

According to the present invention, it is possible to speed up file management processing when handling a large amount of data, and it is possible to improve the processing performance of the entire computer system.

[Brief explanation of the drawing]

Figure 1 is a file configuration diagram created on the auxiliary storage device.
Fig. 2 is a logical configuration diagram of an index file, Fig. 3 is a structural diagram of a conventional index file on an auxiliary storage device, Fig. 4 is a structural diagram of an index file of the present invention, and Fig. 5 is an embodiment of the present invention. It is a processing flowchart.
. Stem 40 Stem S diagram

Claims (1)

[Claims] 1. In a computer system that processes a large amount of data, the data is stored on an auxiliary storage device and read and referenced to the processing device as necessary.
A collection of related data is managed as a file, and an index-only file is created to quickly search for any data within this file, and each node from the root of the tree structure to the leaves is sequentially indexed. , When constructing the index-only file on the auxiliary storage device, nodes at matching levels are set at regular intervals in consideration of the time required for searching the node data on the processing device side. A file management method that is characterized by arranging files with only a few spaces between them.