JPH11219306A - System for storing record of file - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute reorganization processing of a file without suppressing an access to a file. SOLUTION: A record being a retrieval/update object is stored in an objective file. A retrieval file 105 manages to which objective file in files 103, 104, 108 and 109 the record being the retrieval/update object is stored. A storage condition table 102 manages the storage condition of the record in the objective file. In the system, the record being a reorganization object is stored in one objective file and the extraction processing of the record being the reformation object is not required. The objective file is blocked by changing the storage condition and it can temporarily be removed from the object of retrieval, addition and update. Thus, a processing can be executed in parallel to the retrieval/ addition/update of the other objective file during reorganization of the record stored in the objective file. Namely, it is not necessary to suppress an access to the record during reorganization.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file reorganization method, and in particular, can delete stored unnecessary records without stopping the processing even during online processing. Regarding the record storage method.

[0002]

2. Description of the Related Art Generally, records are added, updated, and deleted in a file. In particular, the addition and updating of records are processed sequentially as necessary, but the deletion is rarely processed except for the case where inconvenience occurs due to the presence of the relevant record. Therefore, unnecessary records are stored in the file forever, and the record storage area of the file increases. Further, the physical arrangement on the storage device in which the record is stored may be disturbed, which may cause a decrease in the access performance of the record. Therefore, it is necessary to determine certain conditions and delete unnecessary records. This is called file reorganization. In these reorganizations, all stored records are read out, backed up to another storage device, and then, under certain conditions, records to be stored are extracted and the records are stored again in the original storage device. However, in such a process, if the item that becomes the storage condition is changed during the extraction of the record to be re-stored, it will not be reorganized correctly, so it is necessary to suppress the addition, update, and deletion of records during reorganization was there. Therefore, 2
In a system that operates for four hours (for example, a hospital online system that operates even on holidays and nights), a reorganization time is determined, the system is stopped, and reorganization processing is performed. Further, as an invention aimed at localizing the range of access suppression of records during reorganization, a “database management system and a database management method” (Japanese Patent Laid-Open No.
-13911) is described in the official gazette.

According to the present invention, data is M (M> 1)
The M pieces of data are divided and stored in N (N> M) storage devices, and the M pieces of data are sequentially moved in the N storage devices, so Access only to data is suppressed to achieve localization.

[0004]

SUMMARY OF THE INVENTION It is an object of the present invention to enable a record to be reorganized without inhibiting addition, update and deletion of a record.

[0005]

According to the present invention, a plurality of target files storing records, a search file storing the name of the target file storing the records, and a storage of the target files are provided. It consists of a storage condition table that manages conditions.

According to the present invention, if the record to be reorganized is one,
Are stored in two target files, and the process of extracting the reorganization target records is unnecessary. Further, by changing the storage condition table, the target file is blocked, and search, addition,
Since it can be temporarily removed from the update target, it can be processed in parallel with the search, addition, and update of other target files even during reorganization of the records stored in the target file. That is, there is no need to suppress record access during reorganization.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a file storage system according to the present invention will be described below.

FIG. 1 shows the structure of a file according to the present invention. Reference numeral 101 denotes a search file. 10
2 is a storage condition table. 103, 104, 108, 10
Reference numeral 9 denotes a target file for storing records. In this example, there are four target files, but the number of target files may be any number as long as it is two or more. The search table 105 is
Which target file (1
03, 104, 108, and 109). The storage condition table (102) manages the storage conditions of the target file. In this example, four target files are prepared, and the record update date (107) in the record (106) to be searched / updated is stored as the storage condition.
Is trying to determine what month it is. File 1 stores the August update. File 2 stores the September update, and file 3 stores the October update. Also, the storage condition OF of the file 4
F indicates that the target file is in a blocked state. The condition of the reorganization in this example is that the record update date is 3 months or more (in October, the update in July, and in November, the update date is 8 months).
Monthly update) will be deleted. The target file stores records to be searched and updated.

FIG. 2 shows a flowchart in the case of referring to the record of the target file in the file structure shown in FIG. When the search program specifies the record key of the target file and instructs the record reference of the target file, first, the search file (1
01) A search is made for a record that matches the key in (step 201). It is determined whether or not there is a record of the corresponding key in the search file (101) (step 202). If there is no record, information indicating that there is no corresponding record is returned to the program (step 206), and the process ends (step 207).
If there is a record, the storage condition table (102) is searched for a storage condition that matches the file name of the record (step 203), and the storage condition is determined (step 20).
4) If the record is OFF, that is, if the record is in a state after the reorganization processing, the absence of the record is returned to the program (step 206), and the process ends (step 20).
7). If it is not OFF, reference is made to the record of the target file having the file name of the record of the search file (101) (step 205), the record information is returned to the program, and the process ends (step 207).

FIG. 3 shows a flowchart in the case of updating the record of the target file in the file structure shown in FIG. When the update program designates a record key of the target file and issues a record update instruction of the target file, the search file (101) is searched for a record having the same key using the update instruction key (step 301). ). The presence or absence of the corresponding record in the search file (101) is determined (step 302). If there is no record, the absence of the corresponding record is returned to the update program (step 306), and the process ends (step 31).
1). If there is a corresponding record, the storage condition table (10
After searching for a storage condition that matches the file name of the relevant record from 2) (step 303), the storage condition is determined (step 304). If it is OFF, no record is returned to the update program (step 306), and the process ends. (Step 311). This is already returned to the program (step 306), and the process ends (step 311). This is a state in which record reorganization processing has already been performed and record updating cannot be performed. If it is not OFF, it is determined where the date when the record update request is received from the update program in the storage condition table (102), and the corresponding file name is acquired (step 305). In this example, if the record update date of the record requested to be updated by the program is during September, file 2 is determined, and if it is during October, file 3 is determined. The file name of the corresponding record in the search file is compared with the target file determined in step 305 (step 307). If they match, the record of the target file determined in step 305 is updated (step 310), and the process ends (step 310). Step 311). Step 305: the file name of the corresponding record in the search file
If the target files obtained in step do not match, step 3
A record is added to the target file obtained in step 05 (step 308), the file name of the record of the search file is changed to the target file name obtained in step 305, updated (step 309), and the process is terminated (step 311). .
As a result, the updated record is moved to the target file corresponding to the latest record update date, and the record stored before that becomes an unnecessary record. Unnecessary records are deleted during the file reorganization process. The key is “000-001” for the search file and target file in FIG.
-7 "," 000-002-4 "," 000-003-
FIG. 7 shows a search table and target files when the update processing was performed on October 1 for “1” and “000-004-8” (701, 702, 703, 704, 705, and 70).
6). However, the update process of “000-004-8”
Since there is no corresponding record (step 306), the record is not updated to the target file.

FIG. 4 is a flowchart showing a case where a record is added to a target file in the file structure shown in FIG. When the additional program designates a key and instructs to add a record of the target file, a search is made for a record having the same key from the search file (101) (step 401). It is determined whether or not there is a corresponding record in the search file (step 402). If there is no corresponding record, it is determined where in the storage condition table (102) the date when the record update request is received from the additional program is applicable, The corresponding file name is obtained (step 405).

In this example, if the record update date of the additional record is in September, file 2 is determined, and if the record update date is in October, file 3 is determined. Next, a record is added to the target file obtained in step 405 (step 40).
6) The target file name obtained in step 405 is set as the file name in the record of the search file, and the record is added to the search file (step 407).
The process ends (step 412). The presence or absence of the corresponding record in the search file is determined (step 402). If there is the corresponding record, the storage condition table (102) is searched for a storage condition that matches the file name of the record (step 403), and stored. The condition is determined (step 404). If the storage condition is not OFF, a record duplication error is returned to the update program because the record to be added is already stored in the target file (step 411).
The process ends (step 412). When the storage condition is OFF, that is, when the data has been stored once in the past but has been deleted by the reorganization process. At this time, it is determined where the date when the record update request is received from the additional program in the storage condition table (102), and the corresponding file name is obtained (step 408). Step 408
Add a record to the target file determined in (4)
09), the target file name obtained in step 408 is set as the file name of the record of the search file, the update is performed (step 410), and the process ends (step 41).
2). The keys for the search file and the target file in FIG. 1 are “000-001-7” and “000-004-”.
8 and “000-006-2”, the search table and the target file are shown in FIG.
(801, 802, 803, 804, 805, 8
06). However, “000-001-7” corresponds to step 4
11, a record duplication error occurs.

FIG. 5 shows a flowchart in the case of deleting the record of the target file in the file structure shown in FIG. When the deletion program issues a record deletion instruction, a record matching the key in the search file is searched (step 501). It is determined whether or not the corresponding record exists in the search file (step 50).
2) If there is no corresponding record, a deletion error is returned to the deletion program (step 503), and the process ends (step 508). If there is a corresponding record, after searching the storage condition table (102) for a storage condition that matches the file name of the record (step 504), the storage condition is determined (step 505), and if the storage condition is OFF, it is deleted. The error is returned to the deletion program (step 503), and the process ends (step 508). If the storage condition is not OFF, the record of the target file is deleted (step 50).
6), the corresponding record in the search file is deleted (step 507), and the process ends (step 508).

As described above, the target file is determined in accordance with the storage conditions set in the storage condition table, and records are added, updated, and deleted. As a result, the records of each target file have the same storage conditions. Since this storage condition matches the reorganization condition in advance, all records included in a certain target file are in the same state as extracted by the reorganization condition.

FIG. 6 shows a flow when a reorganization process is executed in the file structure shown in FIG. Date of reorganization (November 1 in this example)
Then, the storage condition of the file 4 is changed from the storage condition table to OFF, that is, the storage condition of the file 4 is changed (step 60).
1). In this example, “the record update date is November” is set. As a result, the target file that has been blocked is set as an access target. Next, the update condition of the file to be reorganized (the file updated in August since it is three months or more from November, that is, file 1) is changed to OFF (step 602). After closing the target file that has been turned off, the target file is initialized (step 60).
3), end (step 604).

[0016]

According to the file record storage method of the present invention, the file reorganization processing and the file access can be performed in parallel, so that the file reorganization processing can be performed without stopping the online system. .

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing the structure of a file according to the present invention.

FIG. 2 is a flowchart showing a record reference procedure in a file storage method.

FIG. 3 is a flowchart showing a record update procedure in a file storage method.

FIG. 4 is a flowchart showing a procedure for adding a record in the file storage method.

FIG. 5 is a flowchart showing a record deletion procedure in the file storage method.

FIG. 6 is a flowchart showing a procedure for reorganizing a file to be deleted in the file storage method.

FIG. 7 is a file after performing an update process on the file of FIG. 1;

FIG. 8 is a file obtained by performing an addition process on the file of FIG. 1;

[Explanation of symbols]

101: Search file, 102: Storage condition table, 103: Target file File 1, 104: Target file File 2, 105: Search table,
106: Record of target file, 107: Record update date of record of target file, 108: Target file File 3, 109: Target file File 4, 109: Target file File 4, 701: Search table after update processing, 702 ... Target file after update processing
File 1, 703: Target file after update processing File 2, 704: Target file after update processing File 3, 705: Target file after update processing File 4,
706: Storage condition table, 801: Search table after addition processing, 802: Target file after addition processing File 1, 803 ... Target file after addition processing File 2, 804 ... Target file after addition processing File 3,
805: Target file after addition processing File 4, 80
6. Storage condition table.

Claims (1)

[Claims] In a system that requires reorganization of files, two or more target files for storing records, and storage conditions that are pairs of storage conditions based on the names of the target files and update dates In a file structure consisting of a table and a search file in which key information of a record and a name of a target file storing the record are paired,
A record storage method, wherein a reorganization condition of a record stored in a target file can be changed by changing at least one item of a search file and a storage condition table.