JPH0317727A - Record input/output control system - Google Patents

Abstract

PURPOSE:To decrease the number of access times to a file and to shorten the processing time by attaining the addition, the change and the deletion of records to a data buffer without having a direct access to an actual file. CONSTITUTION:When a command identification means 13 accepts a file open request, the control is shifted to the file open processing part of a command corresponding process means 14. Thus a log file 26 is opened and the records bearing the relative record Nos 1 - 514 are read out of the file 26 within a data buffer 25 as a buffer No.0. When the means 13 receives a read request for the record of a relative record No.200, the control is shifted to the record read processing part of the means 14. Thus the presence or absence of a designated record is checked within the buffer 25, and a calculation is carried out based on a flow. Consequently, the buffer No.0 is obtained as the record No.200 in the buffer 25. As the buffer No. of the record read in a file open process is equal to 0, the record of the No.200 is read out of the buffer 25 and sent back to a command requester 11.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention is used in the field of small computers to manage record input/output when reorganizing log files where data is frequently input/output. [Conventional technology] A log file is a file that records the contents of records and file updates in chronological order when a certain system or program is operated. Conventionally, when reorganizing a log file containing such a large number of records, the file was accessed each time one record was processed. File editing or relative v4 of this log file,
As shown in FIG. 6, this is a file in which a specified record 62 can be directly read based on the relative record number 6l indicating the number of each record from the beginning of the file. To re-t4m a file, first read the log file one record at a time from the beginning and delete unnecessary records. In order to remove the deleted record area, it is necessary to rearrange the records. Such processing was performed directly on the file by specifying the relative record number of the relative organization file.

[Problem to be solved by the invention]

As mentioned above, a large number of records are written to the log file, so if you directly access and process the file record by record as in the past, the file I
The number of /O operations becomes extremely large, resulting in a long processing time. [Means for Solving the Problems] According to the present invention, ``In reorganizing a log file composed of relative network files, a command identification means for distributing processing according to instructions from a command request source, and a command identification means for distributing processing according to instructions from a command request source, and Command-compatible processing means for processing corresponding files and records, and file management for inputting and outputting records between files and data buffers, and changing the buffer read size to match the file size during input and output. A record management means for checking the existence of a designated record in a data buffer and issuing a buffer replacement request if the record does not exist, and a temporary storage means for storing a set of records read from a file. A new record input/output management method is obtained.

〔Example〕

Next, the present invention will be described in detail with reference to the drawings showing one embodiment of the present invention. Referring to FIG. 1, the record input/output management method 12 according to an embodiment of the present invention includes a command identifying means 13 that allocates processing according to a command specified by a command request source 11, and a command identifying means 13 that allocates processing according to a command specified by a command request source 11, and a command that performs processing corresponding to the command. The information is received from the corresponding processing means 14, the file management means 23, the record management means 24, and the temporary storage means 25 called a data buffer. The command identifying means 13 receives 8
Since a type of command is specified, a processing request is made to the command corresponding processing means 14 corresponding to the command. The command corresponding processing means 14 performs processing corresponding to eight types of commands requested from the command request source 1]. 8
The types of processing are file opening 15, file closing 16, file size acquisition 17. These are record reading 18, record writing 19, record replacement 20, record deletion 21, and commit processing 22. Since these processing units operate independently, each processing unit will be explained.

The file open processing unit 15 opens the log file 26 and reads and stores the data buffer's worth of records from the beginning of the file into the data buffer 25.

The file close processing unit 16 writes the record currently in the data buffer 25 to the log file 26, and then closes the log file 26.

The file size acquisition processing unit 17 uses the log file 26
Find out how many records can be stored in and notify you. The record reading processing unit 18 checks whether the specified record exists in the data buffer 25 at that time, and if it does not exist, reads the range in which the specified record exists into the data buffer 25. Once completed, the specified record is passed to the command request source 11. The record write processing unit 19 checks whether the specified record exists in the data buffer 25 at that time, and if it does not exist, writes the range in which the specified record exists to the data buffer 25.
The specified record is read into the data buffer 25.
Store it in . No writing is done to the log file 26 at this point.

The record replacement processing unit 20 replaces the specified record with the input record, but if the specified record does not exist in the data buffer 25, the range where the specified record exists is replaced with the data. The data is read into the buffer 25 and record replacement is performed on the data buffer.

The record deletion processing unit 21 deletes one of the specified records.
The process of setting a certain code to the byte does not clear all the contents of the record. This processing is also performed on the data buffer 25, so if the designated record is not on the data buffer 25, the range where the designated record exists is read into the data buffer 25.

The commit processing unit 22 saves the set of records in the data buffer 25 at the time the command was requested to the log file 2.
6 is the writing process. The log file 26 is a place where records to be processed are stored, and its size is not constant. That is, there are cases where it is divisible by the data buffer size, as shown in FIG. 2, and cases where it is not divisible by the data buffer size, as shown in FIG. The file management means 23 performs processing for reading and writing a record set from the log file 26 to the data buffer 25. Also, when performing input/output to the file, it manages changes in the buffer size.

Referring to FIG. 4, which shows a flowchart of the buffer size change management method, explanation will be given along the flow. First, in order to know the size of the entire file, file size acquisition processing 41 is performed.

File size acquisition only needs to be done once while this program is running, so if you have done it before, you can omit this process.

Next, it is checked whether the specified relative record number is within the range of the number of records calculated from the obtained file size (step 42). As a result of the check, if it is determined that the designated relative record number is within the range of this file, the buffer size is set (step 43). In setting the buffer size, first set the maximum size of the data buffer, that is, the number of records of n records shown in FIGS. 2 and 3. Normally, n records of this size are read into the buffer. Next, it is checked whether the record set to be read is the m record portion shown in FIG. 3 (step 44).
).

If it is an m record portion, the number of m records is calculated and the preset buffer size is changed to that value (step 45). When inputting/outputting a record to/from the data buffer 25, the record management means 24 checks whether a record with a designated relative record number exists in the data buffer 25 at that time.

The method for checking the existence of a designated record in the data buffer will be explained with reference to FIG. 5, which is a flowchart. First, a buffer number 51 and an in-buffer relative record number 52 are calculated from the designated relative record number. Buffer number 51 is shown in Figures 2 and 3.
As shown in the figure, when all records in the log file 26 are separated by the size to be read into the data buffer 25, this number indicates the buffer number from the beginning of the file. After the calculation, if the relative record number in the buffer is rQj, the calculated buffer number is decreased by "1" to make the relative record number in the buffer the final relative record number in the data buffer (step 53). next,
The buffer number of the record set stored in the data buffer 25 at that point is compared with the calculated buffer number (step 54). This is a record existence check. If the two buffer numbers are the same, the specified record exists in the data buffer 25, and if the buffer numbers are different, it means that it does not exist. The data buffer 25 stores a set of records read from a file and updates each record. This method performs a series of operations based on command instructions from the command request source 11. Here, "Open file"
An example of the operation when a series of command requests such as ``read record'', ``delete record'', and ``close file'' are accepted is shown below.

First, when the command identification means 13 receives a file open request, control is transferred to the file open processing section of the command correspondence processing means 14.

Here, the log file 26 is opened, and relative record number 1 is stored in the data buffer 25 as buffer number O.
Records from to 514 are read from the log file 26. In this case, the buffer record size is 514 (n
= 514), and the size of one record is 128 bytes. Next, when the command identification means 13 receives a record reading request with relative record number 200, control is transferred to the record reading processing section of the command correspondence processing means 14. Here, we first check whether the specified record exists in the buffer. When calculation is performed according to the data buffer existence check flow shown in FIG. 5, the buffer number of relative record number 200 becomes O, and the relative record number in the buffer becomes 200. The buffer number of the record read in file oven processing is O, so
This means that the specified record exists in the data buffer 25. Therefore, the record with buffer relative record number 200 is read from the data buffer 25 and returned to the command request source 11. Next, the command identifying means 13 selects the relative record number 450.
When the record deletion request is received, control is transferred to the record deletion processing section of the command corresponding processing means 14. Here, as in the record reading process, the existence of the specified record in the buffer is checked. When calculated, the buffer number is O, and the relative record number in the buffer is 4.
It will be 50. Therefore, since the specified record exists in the data buffer 25, the data buffer 25
All you have to do is delete the record with relative record number 450 in the buffer. Finally, when the command identifying means 13 accepts the file close request, control is transferred to the file closing processing section of the command corresponding processing means 14. Here, 5' of buffer number O on the data buffer 25 at that time. 4
All records are written to the log file 26 and the log file 26 is closed. [Effects of the Invention] As explained above, according to the present invention, records can be added to and from the data buffer without directly accessing the actual file. Changes and deletions can now be made, the number of accesses to files can be reduced, and processing time can be shortened.

[Brief explanation of drawings]

FIG. 1 is a schematic diagram of the structure of one embodiment of the present invention, and FIG.
A diagram showing an example where the file management means in the figure is divisible by the buffer size in the buffer operation when reading records from the file to the data buffer. Fig. 4 is a flowchart showing the processing when checking the buffer size of this embodiment, and Fig. 5 specifies this embodiment. A flowchart showing the process of checking the existence of a record in the data buffer, and FIG. 6 is a schematic diagram of the record structure of a relative file. 11: Command request source, 12: Record input/output management method, 13: Command identification means, 14
: Command corresponding processing means, 15: File open processing unit, 16: File close processing unit, 17: File size acquisition processing unit, 18: Record reading processing unit, l9: Record writing processing unit, 20: Record replacement processing section, 21: record deletion processing section, 22: commit processing section, 23: file management means, 24: record management means, 25: temporary storage means (data buffer),
26: Log file.

Claims (1)

[Claims] In reorganizing log files composed of relatively organized files, there is a command identification means that distributes processing according to instructions from the command request source, a command corresponding processing means that processes files and records corresponding to each command, and files and data. A file management means that inputs and outputs records to and from the buffer and changes the buffer read size to match the file size during input and output, and checks the data buffer for the existence of a specified record and confirms that it does not exist. A record input/output management method comprising: a record management means for issuing a buffer replacement request; and a temporary storage means for storing a set of records read from a file.