JPS61145658A - Index order editing file device - Google Patents

Abstract

PURPOSE:To prevent overflow of a normal record to the utmost by writing an added record while deleting a record to be deleted and compressing it when the added record is generated. CONSTITUTION:When an additional record A1 to be inserted at a position next to a record Rj of a block Bi is generated, records after the record D1 to be deleted are shifted to the right by one each, the record A1 is written in a location where the record D1 to be deleted existed, a record Dp to be deleted overflowed from the right end of the block Bi is deleted and compressed. When the additional record A2 to be inserted between the records R1 and R2 is produced, records after the record R2 are shifted to the right by one, the additional record A2 is inserted to the vacant location and a record Dp-1 to be deleted is deleted. Every time the additional record is generated, a record Dk to be deleted is deleted and compressed and the additional record is written.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an indexed sequential file device that aims to shorten access time and reduce the frequency of file reorganization.

[Prior Art] As is well known, in a direct access device such as a magnetic disk device, records can be processed in the order of keys [sequential access], and C that directly processes only a specific record by specifying a key. Indexed sequential files, which can also be accessed (randomly accessed), are widely used.

FIG. 4(a) is a conceptual diagram showing an example of the structure of a conventional index sequential organization file. In this figure, 1 is the file label indicating the file name, 2 is the ninth index area for searching records, and 3 is the block block (1=1.2...
m), and the data area S4 for storing records is an overflow area for storing records overflowing from the data area 3. Here, the block B1 generally includes a control unit C and a record Rj (j=1 * 2) that actually stores data, as shown in FIG.
...n) and a free area S. It was,
A key is assigned to each record Rj.

Rj is stored in the data area 3 in ascending order of keys. The key of the record stored last in each block is recorded in the index area 2, forming an index.

In this typical configuration, when an additional record occurs,
Additional records are inserted using ELL as follows.

■ Search the index area and search for block B1 to be added.

■ Find an insertion position where the key of the record in the Todoroki block Bi is kept at Ascension 1@, and write the records after the insertion position one by one to the right to create a free space), and write the above additional records. Write.

■ As a result of inserting additional records, the records overflowing from the right end of block B1 are moved to free space S or deletion flow) (
(This will be discussed later), the record is deleted.

■ If the overflow nine records are normal records, ′
It is accommodated in the overflow area 4.

In this way, in an indexed sequential file, records are always stored in ascending order of keys, allowing both sequential and random access, and efficient file management.

[Problem that the invention seeks to solve]

By the way, in the above-mentioned conventional indexed sequential file, the data area 3 and the automatic flow area 4 are usually located at independent locations.

Therefore, as the number of records in the overflow area 4 increases, the number of alternate accesses between the data area 3 and the oats 5-flow area 4 increases, resulting in an increase in extra seek time and an increase in access time. There was a problem. Furthermore, the records in the flow area 4 are not blocked, but are related by links, which also causes delays in processing.

In order to prevent such a decrease in processing speed, reorganize the file so that each flow V is arranged in ascending order of keys.
It must be moved to data area 3. However, with this special file reorganization program, not only can you register in the file, but you can't reference or output the file at all. There was a problem.

In order to avoid these problems, methods are used to reserve a large amount of free space or dummy records in advance, but these methods have the following drawbacks.

(No. 11 free area S is generally all blocks Bl to BnJ
Since this is commonly reserved, it is not possible to reserve more records only in places where additional records are likely to be concentrated. As a result,
In a block with few additional records, the free area S remains unused and the usage efficiency within the block decreases.

(2) In the method of securing dummy records (that is, records that do not contain meaningful data), the dummy records are treated in the same way as normal records when read, so the next record is read by the processing program and the next record is a dummy record. It is necessary to determine whether or not it is a record before processing, which makes the processing complicated.

In addition, as shown in Figure 5, there is a method of separately preparing an existing record file and an additional record file, and merging them at a predetermined period to create a prayerful existing record file. As shown in Figure 6, a possible method is to change from an index sequential file to another file such as a direct file and manage them separately.

However, as shown in the table, the former method has problems in that centralized file management is difficult, and the latter method restricts random access and sequential access, which are characteristics of indexed sequential files.

This invention attempts to solve the above-mentioned problems, particularly the increase in access time and the increase in the frequency of file reorganization.

Means for Solving Problem c] In order to solve the above problem, the present invention eliminates deleted records that are skipped during reading and are deleted and compressed when inserting additional records, which are expected to be concentrated in additional records. The feature is that the data area is reserved in advance.

[Effect]

According to the above means, when an additional record is generated, the deleted record is deleted and compressed and the additional record is written, so that overflow of normal records can be prevented as much as possible. Furthermore, since deleted records are skipped when read and are not read into the processing program, they do not need to be processed as frequently as dummy records.

〔Example〕

Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a basic configuration diagram showing the configuration of one practical example of the present invention. In this figure, 11 is a storage device such as a magnetic disk, 12 is a writing means, and 13 is a reading means. Each area similar to that shown in FIG. 4 is secured in the storage device 11, and the record Rj in block Bi has a flag section F, a key section, and a data section, as shown in FIG. It is composed of DT. In this flag section F, when the record is a deleted record, FF (1g notation) is written, and it is marked I to distinguish it from other records.

Then, in block B1 where additional records are expected to be concentrated, the 3rd v! J(a)l A deletion record Dk (k=1.2-"'El) as shown is secured in advance after record Rj. Note that block B of this embodiment
1ζ No free space is taken.

In such a configuration, record Rj of block B1
If an additional record AI to be inserted at the next position occurs, the records after deleted record R2 are written one by one to the right, and additional record A1 is written at the location where the deleted record DI was, and the block The nine deleted records DXI overflowing from the right end of B1 are deleted and compressed (Figure 3).

Next, if an additional record A2 is to be inserted between records R1 and R2, the records after record R2 are written one by one to the right, leaving the empty space (where record R2 was stored). When the additional record A2 is inserted in the location], the deleted record Dp-□ is deleted (Fig. 3 (9)).In this way, each time an additional record is generated, the deleted record Dk is deleted and compressed, and the deleted record Dk is deleted and compressed. Records are written. Therefore, deletion records D0 to D
Records that overflow from block B1 and are transferred to overflow area 4 until all p are deleted and compressed.
will not occur.

Additional data written as described above.

It can be read out in the same way as before. Further, the remaining deletion records are skipped as determined by the flag F, and are therefore not read into the processing program.

When the file is reorganized after a considerable amount of additional records have been written, the remaining deleted records are deleted and other records are stored in the data area 2 in ascending order of keys.

According to this embodiment, overflow does not occur as long as there are deletion records Dk, so it is possible to suppress record rewriting to the overflow area 4. Therefore, it is possible to prevent an increase in access time and to reduce the frequency of file reorganization. As a result, file usage efficiency can be improved.

In addition, any number of deletion records D1 to Dp (in ft and less than n) can be secured in any block B1 by key operation, so less space is wasted (when the area is free). It's over.

Furthermore, unlike dummy records, deleted records are not read into the processing program, so there is no need to judge whether or not they are dummy records, which has the advantage of simplifying processing.

〔Effect of the invention〕

As explained above, the present invention secures in advance deleted records that are skipped when reading and deleted and compressed when inserting additional records in a data area where additional records are expected to be concentrated, thereby achieving the following effects. can be played.

(It is possible to suppress the increase in access time due to the addition of 11 records.

(2) The method of securing free space and the efficiency of file batching can be improved.

(3) The frequency of file reorganization can be reduced.

[Brief explanation of drawings]

FIG. 1 is a basic configuration diagram showing the configuration of an embodiment of the present invention;
FIG. 2 is a conceptual diagram showing the structure of blocks and records in the same embodiment; FIG. 3 is a conceptual diagram showing the insertion state of additional records; FIG. 4 is a conceptual diagram showing the structure of a conventional indexed sequential file; 5 and 6 are diagrams for explaining the conventional file structure. 3...7''-vine area 4...overflow area 11...storage device 12...writing means 13...reading means AI , A2...Additional record Bi (1=1.
2...m)...Block Dk (k=1.2・
...p)...Delete record K...Key part Rj (j=1.2...n)...Record S...
·····Free space. (4) Figure 4

Claims (1)

[Claims] A storage device having a data area and an overflow area consisting of a plurality of blocks, and a write function that writes records to which keys are attached to the blocks in the order of key numbers, and writes records that overflow from the blocks during the write operation to the overflow area. and reading means for sequentially or randomly reading desired records using the key as an index, wherein the reading means skips reading and the writing means deletes/compresses. An indexed sequential file device characterized in that records are secured in advance in a data area where additional records are expected to be concentrated.