JPH07104758B2 - Record processing method - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a record processing system, and more particularly, to processing of variable-length records in blocks in a data processing system.

2. Description of the Related Art Conventionally, in this type of processing method for variable-length blocked records, in response to a record deletion request from a program, a record for which a deletion request was made without making the area where the record existed free It moved logically later records so that there was no free space between records in the block.

In addition, when inserting a record, in order to make the logical order and the physical order of the records in the block correspond, after searching the block in which the record should be inserted, the records are searched in the logical order of the records and inserted. When the logical position to be inserted is found, the physical position of the record that is logically later than the record to be inserted is moved to secure a continuous empty area to be inserted, and a new record is inserted there.

Also, the program may store the address of a record that has been accessed once and then access it again. However, as described above, deleting or inserting a record changes the address of the record storing those addresses. Sometimes. At that time, if the information is not notified to the program, the operation may become illegal when an access request is made again by the address.

In the conventional processing of variable-length records that have been blocked as described above, a record move occurs for each request to insert or delete a record, and unnecessary record moves, the number of record moves, and record Has the drawback of increasing the amount of movement.

For example, in FIG. 10, when the records A, B, and C existing in a certain block BL1 are deleted in order and all the records in the block BL1 are deleted, in the conventional technique, when the record A is deleted, the records are deleted. When B and C were moved and record B was deleted, it was necessary to move record C. For this reason, there is a drawback in that an originally unnecessary record movement occurs.

Further, in FIG. 11, when records D, F, G exist in a certain block BL2 and a record E is inserted in the block BL2, the records F, G are moved to keep the logical order of the records, It was necessary to insert record E between D and record F. Here, assuming that the record lengths of the records D, E, F, and G are 50, 40, 120, and 120 bytes, respectively, the records F and G must be moved by 40 bytes to insert the record E. First, in order not to destroy the image of the original record, first move the last 40 bytes of the record G. next,
The process of moving the next 40 bytes to the free area is 6
The record E is inserted, the area for inserting the record E is secured, and the record E is inserted. In other words, there is a drawback that the number of times of movement increases because it is necessary to perform the movement process for the number of times obtained by dividing the sum of all record lengths to be moved by the amount of area that must be free.

In addition, in the related art, it is possible to delete or insert a record,
As shown in Fig. 11 and Fig. 11, when the physical positions of the records B, C, F, G are moved and the program has already stored the addresses of those records, the change of those addresses is made to the program. It had the drawback of having to notify.

OBJECT OF THE INVENTION The present invention has been made to eliminate the above-mentioned drawbacks, and an object of the present invention is to reduce the movement of records within a block as much as possible and to perform processing such as record deletion and insertion. It is to provide a record processing method that does not require the change of the record address of.

Configuration of the Invention The method of the present invention is a record processing method in a data processing system in which variable-length data records are divided into blocks and stored in a file in accordance with an access request from a program. A record descriptor storage area is provided at each location corresponding to each record in this data block, and in each of these record descriptors, corresponding record storage location information indicating the storage location in the storage area of the corresponding record, and the storage location Record status information indicating the presence / absence of a corresponding record in the area, logical order information indicating which record is logically the record next to the corresponding record in the data block, and record length information of the corresponding record are included. In addition, when deleting a record, When inserting the record by rewriting the record state information of the record descriptor to the deleted state and logically rewriting the logical order information of the record descriptor of the previous record to the order excluding this record to be deleted, , A record descriptor in which a record to be inserted is stored in an empty area of a data block, and record storage position information, record status information, logical order information, and record length information corresponding to the record are set as a record descriptor. If there is no free space in the storage area, it is created at the end of the record descriptor storage area, and the logical order information of the record descriptor of the record logically preceding the record includes this insertion target record. Rewrite in sequence and access the record with the data block number and the sequential number of the record descriptor in the data block It is characterized by doing so.

Embodiments of the present invention will be described in detail with reference to the drawings.

With reference to FIG. 1, an embodiment of the present invention comprises an access request analysis means 1, a block analysis means 2, a record search means 3, a record transfer means 4, a record descriptor maintenance means 5,
It is composed of a free area discriminating means 6, a record moving means 7, a block reading means 8, a block writing means 9, a block storage area 10, a file 11, a program record processing area 12 and a program access request 13.

Referring to FIG. 2, the block for storing the variable length record includes a block header 21, a plurality of records 22, and an empty area.
It consists of 23 and multiple record descriptors 24.
Record descriptor 24 exists corresponding to record 22, and each record descriptor RD1, RD2, RD3 is record R1, R2, R3, respectively.
It corresponds to.

In each record descriptor, record status information, record length information, record storage position information in the block,
It contains information for maintaining the logical order of records.
The record descriptor 24 in the block is sequentially numbered in the order of RD1, RD2, and RD3, and the relationship information of the sequential numbers is stored in the record descriptor 24, so that the logical order of the records is maintained.

In the block header 21, the total free area amount information in the block, the continuous free area relationship position information in the block, the information for obtaining the logically first record in the block, and for maintaining the logical order of the block Information and are included. The information for obtaining the first record in the block is represented by the sequential number of the record descriptor 24.

Next, the operation of the present invention will be described in detail with reference to the drawings.

First, the operation when there is a record reference request from the program will be described with reference to the operation of each means in FIG.
This will be described in detail with reference to the drawings, FIG. 4 and FIG.

The access request analysis means 1 shown in FIG. 1 performs the steps shown in FIG.
At 31 to 33, the access request 13 from the program is input, the request is analyzed, and the normality of the request is identified. If there is an abnormality here, the error processing of step 37 is performed, and the access request ends abnormally.

The block analysis means 2 in FIG. 1 determines the block number of the request record in step 34 in FIG. The block reading means 8 reads the blocks from the file 11 into the block storage area 10 according to the block numbers in steps 35 to 36. Here, the block number is a number sequentially assigned to each block in the file. The blocks read into the block storage area 10 have the structure shown in FIG. 2 as described above.

The record search means 3 of FIG. 1 uses the logically first record descriptor 24 in the block in step 41 of FIG.
(There is a case where the nth record is physically deleted from the 1st record by the previous access request, so the n + 1th record is found.) . In steps 42 to 44, the position of the record in the block is obtained based on the obtained sequential number, the logically first record is found, and it is judged whether or not the record matches the access request.
If they do not match, the request record is found by the procedure of steps 45, 46, 42 to 44. If they match, the record transfer means 4 in FIG. 1 executes the steps in FIG.
At 51 to 52, the record is transferred to the program record processing area, and the address of the record is notified to the program. In step 53, the program is notified of the end.

This record address is the block number and record descriptor
It is expressed by 24 sequential numbers. As a result, even if a record is moved within the block, only the record storage position information in the record descriptor 24 is changed, and the address of the record stored by the program does not need to be changed.

Next, the operation when there is a record deletion request from the program will be described with reference to FIG.
This will be described in detail with reference to the drawings, FIG. 4 and FIG.

The input of the access request, the analysis and the reading of the block are as shown in FIG. 3, the search of the record is as shown in FIG. 4, and the details of the operation are the same as those explained in the case of the record reference request. .

When the record is found, the record descriptor maintenance means 5 in FIG. 1 updates the record state in the record descriptor 24 to the deleted record state in step 61 in FIG. 6, and in the block header 21 in step 62. Add the record length to all free areas in the block. Next, at step 63, the information for maintaining the logical order of the records in the record descriptor 24 of the record logically before and after this record is updated so that the record having the deletion request cannot be accessed in the logical order. In step 64, the record descriptor 24 of this record is made available for reuse. Then, the block writing means 9 shown in FIG. 1 writes the block in step 65. In step 66, the program is notified of the end.

Next, the operation when there is a record insertion request from the program will be described with reference to the third operation while explaining the operation of each means in FIG.
This will be described in detail with reference to FIGS. 7, 7 and 8.

The input of the access request, the analysis and the reading of the block for inserting the record are as shown in FIG. 3, and the details of the operation are the same as those described for the record reference request.

When a block into which a record is to be inserted is found, the free space discriminating means 6 in FIG. 1 discriminates in step 71 in FIG. 7 whether or not there is a free space required to insert a record in the block, and it exists. If not, steps 72-73
Search another block with and read the block. Step
In step 74, it is determined again whether or not there is free space required to insert a record in that block.
Repeat steps 72-73. If the newly read block has enough free space to insert a record, step
At 75, information for maintaining the logical order of blocks in the block header is updated.

Next, in step 81 of FIG. 8, the free space discriminating means 6 discriminates whether or not there is a continuous free space enough to insert a record in the block, and if so, the record descriptor maintenance means 5 In steps 82 to 83, the total free area amount information and the continuous free area start position information in the block in the block header are updated, the record descriptor 24 is created, and this record can be searched in logical order. , Update the information for keeping the logical order of the records in the record descriptor 24 of the record logically before and after this record.

Next, the record transfer means 4 in FIG. 1 transfers the records from the program record processing area 12 to the portion for storing the records in the block in step 84. The block writing means 9 writes the blocks existing in the block storage area 10 to the file 11 in step 85. Step 86
At, the program is notified of the end.

In step 81, if there is not enough continuous free area for inserting a record in the block, the record is moved within the block to create a continuous free area.

The record moving means 7 shown in FIG.
At 89, record descriptors for all records in the block
24 are retrieved, sorted in the order of record storage positions, the records are packed in a block so that there is no free space between records, the records are moved in the sorted order, and the records in the record descriptor 24 of each record Update the storage location information. After moving all the records, update the continuous free space start position information in the record header 21,
After that, the processes of steps 82 to 86 are performed.

Although one embodiment of the present invention has been described in detail above, the record descriptor and the block head are not limited to the positions shown in this embodiment.

The record descriptor includes record status information, record length information, record storage position information, and information for maintaining the logical order of the records. As shown in FIG. It is also possible to divide the descriptor into two parts such as a record descriptor 25 and a record descriptor 24. In the case of FIG. 9, the record storage position information may be put in the record descriptor 24, and the others may be put in the record descriptor 25.

As described above, according to the present invention, regarding the processing of variable-length records in blocks, a record descriptor including a record storage position, a record length, and record status information is provided at a fixed position in the block. , Only update the record status information of the record descriptor in response to the deletion request from the program, and based on the record storage position information in the record descriptor only when there is not enough continuous free space for inserting the record when inserting the record. Then, the record descriptors are rearranged in the order of the record storage positions so that the record movement amount and the number of movements are minimized, and the records are moved in the same block storage area according to the order to secure a continuous free area. Then, by inserting the record so that the logical order of the record is guaranteed by the record descriptor, Even if a record needs to be moved, the number of times of movement and the amount of movement can be minimized.

Also, if the program stores the address of a record that it has accessed once, it becomes unnecessary to change the address of the already stored record by inserting or deleting another record by having the above record descriptor. There is an effect.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a blocked variable-length record processing method of the present invention, FIG. 2 is a diagram showing a block configuration example including variable-length records, and FIGS. 3 to 8 are records. FIG. 9 is a flow chart of processing, FIG. 9 is a diagram showing another example of the configuration of a block including variable-length records, and FIGS. 10 and 11 are diagrams showing an example of a conventional data processing method. Description of code of main part 1 ... Access request analysis means 2 ... Block analysis means 3 ... Record search means 5 ... Record descriptor maintenance means 6 ... Free area determination means 10 ... Block storage area 11 ... File 13 …… Program access request

Claims (1)

[Claims] 1. A record processing method in a data processing system in which variable-length data records are divided into blocks and stored in a file in accordance with an access request from a program, wherein the data is stored in a predetermined location of a data block. A record descriptor storage area is provided for each record in the block, and in each of these record descriptors, corresponding record storage position information indicating the storage position of the corresponding record in the storage area, and corresponding record in the storage area. Record status information indicating the presence or absence of the record, logical order information indicating which record is logically the record next to the corresponding record in the data block, and record length information of the corresponding record. Of the record descriptor of the record to be deleted when deleting Rewrite the code state information in the deleted state,
When rewriting the logical order information of the record descriptor of the record logically preceding the record to be deleted to the order excluding the record to be deleted and inserting the record, store the record to be inserted in the free area of the data block. , A record descriptor in which the record storage position information corresponding to the record, the record status information, the logical order information, and the record length information are set in an empty area in the record descriptor storage area, if the empty area does not exist. Created at the end of the record descriptor storage area, rewrite the logical order information of the record descriptor of the logically previous record to the order that includes this record to be inserted, and set the record to the data block number and the data A record processing method characterized in that it is accessed by the sequential number of the record descriptor in the block.