JPH096653A - Data base checking information processor - Google Patents

Abstract

PURPOSE: To provide an information processor which can check a data base without stopping the addition/updating processing of the data base by displaying an updating mode in response to every record store address in a period when the information on an index part and a data part are updated and preventing a correspondence relation check program from checking the correspondence relation of the corresponding record store addresses. CONSTITUTION: When the information stored in a data base is updated, a data base management program 3 changes the store address of the relative data record of a correspondence relation check table 7 to an updating state from an initial state and then resets the store address in the initial state when the program 3 finishes its operation. A correspondence relation check program 1 refers to the states of both index and data records for every store address of the table 7 and reflects the reference results of both record states on the reference results stored in the table 7. Then it is checked whether the correspondence relation is correct or not between an index key and the data record when the updating state of the store address is reset in the initial state.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing apparatus for checking a database, and more particularly to an information processing apparatus for checking a database in parallel with a process of updating the database according to business.

[0002]

2. Description of the Related Art Generally, a data file or database is composed of an index part and a data part. The index part stores a correspondence relationship between an index key and a storage address of a data record, and the data part is linked by this storage address. The data record accessed by this index key is stored in the storage location.
The program that manages the data file or database searches the index section according to the request of the business program, reads the record storage address corresponding to the specified index key, and based on this record storage address, the purpose in the data section. Access the data record of. When adding a new data record to the data part, the database management program searches the free area of the index part, searches the data part for an unused record storage address, stores the index record in the index part, and then stores the data. Store additional data records in the department.

Databases are generally constructed, accessed, and updated as described above, but the index keys and data records are not always stored in correct correspondence. For example, if the database is physically destroyed, or if incorrect information is stored in the database due to a bug in the database management program, the record storage address corresponding to the index key does not exist, or conversely, the index record exists. However, there may be an inconvenient state in which a single data record exists. For this reason, a program for checking the correspondence between all index keys in the database and data records is created and checked.

Conventionally, the business program is stopped, and the above-mentioned correspondence check program is run in a state where the database is not accessed by others, and the database is checked. However, there is a strong demand to operate the business programs that access the database for 24 hours, and it is difficult to stop all the business programs that access the database and run this correspondence check program. .

Note that, as a technique related to this type, for example, Japanese Patent Laid-Open No. 1-217652, "Simultaneous file update avoidance method", Japanese Patent Laid-Open No. 2-48746, "Exclusive control method", For example, there is a database management system of JP-A-5-21678.

[0006]

Considering the case of adding a data record by the database management program, the additional data record is stored in the data part after the index record is stored in the index part. Therefore, the correspondence between the index key and the data record during this period temporarily becomes inconsistent. Since the conventional correspondence check program does not consider parallel processing with the database management program, running both in parallel causes the correspondence check program to catch this temporary inconsistent state and There is a risk that it will be judged as abnormal. Even so, it is not allowed to stop the business program and run the correspondence program as described above.

It is an object of the present invention to check the database without stopping the database adding / updating process.

[0008]

According to the present invention, during the period in which the database management program updates the information in the index section and the data section, the update is displayed corresponding to each record storage address, and this update is displayed. The information processing apparatus is characterized in that the correspondence check program avoids the correspondence check for the corresponding record storage address during a certain period and performs the correspondence check for other record storage addresses.

[0009]

Since the correspondence check program avoids the correspondence check for the record storage address that is being updated, it is possible to avoid the above-described check for the temporary inconsistency period. If a record storage address that has already been checked is excluded from the check target, and if such a check is tried several times, the record storage address that has avoided the correspondence check once is also checked at the next opportunity.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram of an apparatus for checking a data base according to this embodiment. The database 4 is composed of an index section 5 in which a correspondence table between index keys and data storage addresses is stored, and a data section 6 in which data records are stored. The database management program 3 is a program that controls operations such as reference, update, addition, and deletion of data records in the database 4. The business program 2 is a program that issues a request for operating a data record to the database management program 3. Correspondence check program 1
Is a program for checking whether the correspondence between the index key and the data record is normal, and refers to the index record and the data record on the database 4. The correspondence check table 7 stores the update status of the information in the database 4 for each storage address of the data record, the reference result of the index record and the data record, the check result, etc., and is stored by the correspondence check program 1 and the database management program 3. Referenced and updated. The device of this embodiment can be realized by an information processing device. The correspondence check program 1, the business program 2, and the database management program 3 are programs stored in the storage device of this information processing device and executed by the information processing device. The database 4 and the correspondence check table 7 are data stored in the storage device (main storage device or external storage device) of this information processing apparatus.

When the business program 2 requests the data record of the database 4 such as reference, update, addition, deletion, etc., the database management program 3 refers to the index section 5 and the data section 6 to perform the requested operation. I do. When this operation updates the information in the database 4, the database management program 3 changes the storage address of the related data record on the correspondence check table 7 from the initial state to the updated state, and when the operation is finished, Returns the update state to the initial state. The correspondence check program 1 refers to the state of the index record and the data record for each storage address on the correspondence check table 7, and reflects the result in the reference result in the correspondence check table 7. The correspondence check program 1 uses the storage address where the reference results of the index record and the data record are stored,
When the update state of the storage address is the initial state, it is checked whether the correspondence between the index key and the data record is normal, and the result is stored in the check result column on the correspondence check table 7.

FIG. 2A shows the data structure of the index section 5. The index key 17 is an index key for searching a data record. The record storage address 18 indicates a storage address on the data section 6 of the data record corresponding to each index key 17, and the address is designated by a physical address such as a cylinder number, a track number, or a sector number of the magnetic disk device. It may be specified by a logical address such as a page number or a relative byte address. FIG. 2B is a diagram showing the correspondence between the record storage address 21, the usage state 22 and the data record 23 for the data part 6. The record storage address 21 is all addresses on the data section 6 corresponding to the record storage address 18. The usage state 22 is a flag indicating whether actual data is stored (in use) or not stored (unused) in the data record area. The usage state 22 may be prefixed to the data record as control information, or may be identified by whether the first character of the data record is a deletion code (unused) or not a deletion code (in use). Alternatively, it may be set in the form of a bit map corresponding to the record storage address 21 separately from the data record. Data record 2
3 may be a physical record designated by a sector number, a page number or the like, or a logical record designated by a relative byte address or the like. In any case, the index section 5 can be searched based on the index key, and the data section 6 can be searched based on the obtained record storage address to access the data record including the target physical record or logical record. .

FIG. 3 is a diagram showing the data structure of the correspondence check table 7. Record storage address 10
Is the same as the record storage address 18 and the record storage address 21. Update status 11 is database 4
This is a flag indicating whether the information inside is being updated or not being updated (initial), and the initial state is set to “initial”. For the index presence / absence 12, the index key corresponding to the record storage address exists (present) or does not exist (none).
Is a flag indicating whether or not, and the initial state is “none”. The index key 13 is an area for storing the contents of the index key when the index key exists. The data part usage state 14 is a flag indicating whether the usage state 22 corresponding to the record storage address is “in use” or “unused”, and the initial state is “unused”. Check target 1
Reference numeral 5 is a flag indicating whether or not the correspondence between the index key and the data record is normal (target) or completed (non-target), and the initial state is “target”. The check result 16 is a flag that sets the result of the check of the correspondence between the index presence / absence 12 corresponding to the record storage address and the data part usage state 14. Update status 11 is database management program 3
And is referred to by the correspondence check program 1. The index presence / absence 12 to the check result 16 are updated and referred to by the correspondence check program 1.

FIG. 4 is a flow chart showing a procedure of processing of the database management program 3 for adding a data record. When there is a request from the business program 2 to add a data record to the database 4, the database management program 3 searches for an empty area (entry) in the index section 5 (reading the index section: step 41). Next, the usage state 2 of the data section 6
2 searches for a record storage address corresponding to "unused" (reading of data part: step 42).
Next, the update status 11 on the correspondence check table 7 corresponding to this address is changed to "update in progress" (step 43). Next, in the obtained free area on the index section 5, the index key 17 and the obtained record storage address 18
Is stored (writing of the index part: step 44). Next, the usage state of the obtained record storage address is changed to "in use" and the data record is stored (writing of data part: step 45). Finally, the update status 1 on the correspondence check table 7 corresponding to this address
1 is changed to "initial" (step 46). For example, when adding a data record whose index key is “003”, the index key “00
An index record indicating the correspondence between “3” and the address “# 4” is added, the usage state 22 corresponding to the address “# 4” of the data section 6 is changed to “in use”, and the corresponding data record 23 is stored. It

When a deletion request for a data record on the database 4 is issued from the business program 2, the database management program 3 searches the index section 5 based on the specified index key, and the record storage address corresponding to this index key. And the update status 1 on the correspondence check table 7 corresponding to this address
Change 1 to "Updating". Next, the record storage address corresponding to the designated index key on the index section 5 is deleted, and the usage state 22 of the data section 6 corresponding to the obtained record storage address is changed to "unused".
Finally, the update state 11 on the correspondence check table 7 corresponding to this address is changed to "initial". If the size of the data record changes before and after updating in the case of updating the data record, the record before the update will be deleted and then the record after the update will be added. After that, the record addition operation will be performed.

FIG. 5 is a diagram showing a schematic procedure of processing of the correspondence check program 1. First, referring to the index section 5, the index presence / absence 12 and the index key 13 on the correspondence check table 7 are set (step 51). Details are shown in FIG. Next, referring to the data part 6, the data part usage state 14 on the correspondence check table 7 is set (step 52). The details are shown in FIG. Whether or not the next index is set 1
2. Based on the index key 13 and the data part usage state 14, the correspondence between the index key and the data record is checked (step 53). The details are shown in Figure 8.
a, FIG. 8b and FIG. 8c. Finally, the check result is edited and output to an output device such as a printer (not shown) (step 54).

FIG. 6 is a flow chart showing the flow of processing of the correspondence check program 1 for setting the index presence / absence 12 and the index key 13 on the correspondence check table 7. The index records of the index section 5 are read in the arrangement order, and the record storage address corresponding to the index key is obtained (step 6
1). Next, the index presence / absence 12 on the correspondence check table 7 corresponding to this record storage address is set to "present" (step 62). Next, the read index key is stored in the index key 13 corresponding to the record storage address (step 63). If the end of the index section 5 has not been reached (step 64N
O), and returns to step 61 to repeat the above process. If the end of the index unit 5 is reached (step 64YE
S), the process ends. At the end of the process, if there is an index key corresponding to the record storage address on the correspondence check table 7, whether or not there is an index 12
Is set to “present”, and if the index key does not exist, the index presence / absence 12 corresponding to the record storage address remains the initial value “absent”. The process shown in FIG. 6 refers to only the index section 5 in the database 4, and even if the database management program 3 updates the index record, the temporary mismatch with the data record is checked for the correspondence relationship. Does not affect Program 1.

FIG. 7 is a flow chart showing the flow of processing of the correspondence check program 1 for setting the data part usage state 14 on the correspondence check table 7. The record storage addresses 10 registered in the correspondence check table 7 are sequentially referred to, and the usage state 22 of the data section 6 is read for each record storage address (step 7).
1). Next, according to this result, the data part usage state 14 on the correspondence check table 7 is set to "in use" or "unused" for the record storage address (step 72). If the end of the data part 6, that is, the end of the correspondence check table 7 has not been reached (step 7)
3NO), and returns to step 71 to repeat the above processing.
If the end of the data section 6 has been reached (step 73YE
S), the process ends. The process shown in FIG. 7 is a process for referencing only the data part 6 in the database 4, and even if the database management program 3 updates the data record, the temporary mismatch with the index record causes the correspondence check program. Does not affect 1.

FIG. 8a is a flow chart showing the processing flow of the correspondence check program 1 for checking the correspondence between the index key and the data record. For each record storage address 10 on the correspondence check table 7, if the check target 15 is “target” (step 81 YES), the index presence / absence 12 is “present” (step 82 present), and the data part usage state 14 Is "in use" (step 83 in use), the check result 16 is set to "normal" (step 8).
4). If the index presence / absence 12 is "none" (no step 82) and the data part usage state 14 is "unused" (step 85 unused), check result 1
6 is set to "normal" (step 86). If the index presence / absence 12 is “present” (step 82 present) and the data part usage state 14 is “unused” (step 83 unused), the process shown in FIG. 8b is performed (step 87). If the index presence / absence 12 is "none" (no step 82) and the data part usage state 14 is "in use" (step 85 in use), the process shown in FIG. 8c is performed (step 88). After setting a value to the check result 16, the check target 15 is set to "not applicable" (step 89), and if the end of the correspondence check table 7 has not been reached (step 90 NO), step 8
Returning to 1, the above processing is repeated for the next record storage address. If the check target 15 is "non-target" (NO in step 81), the process from step 82 to step 89 has already been completed for the record storage address, so this process is skipped and step 9 is executed.
Go to 0. If the end of the correspondence check table 7 has been reached (YES in step 90) and the correspondence check table 7 has not been scanned a predetermined number of times (step 91 NO), the process returns to step 81 and the first record of the correspondence check table 7 is returned. The above processing is repeated from the storage address. After scanning the correspondence check table 7 a prescribed number of times (step 91YE
S), the process ends. That is, if the record storage address in the "updating" state still remains after repeating a predetermined number of times, the processing is aborted as an exceptional case.

FIG. 8b is a flow chart showing an expanded process of step 87. Index key 13
To search the index section 5 (step 9
3). If the index does not exist (step 94N
O), the check result 16 is set to "normal" (step 95), and the process goes to step 89. If the index exists (YES in step 94) and the record storage address corresponding to the index key 13 matches the record storage address currently being checked (YES in step 96), the check result 16 is set to "abnormal" (step 97). ,
Go to step 89. If the index exists (YES in step 94) and does not match the record storage address corresponding to the index key 13 (step 96N).
O), the check result 16 is set to "normal" (step 98), and the process goes to step 89.

FIG. 8c is an expanded flow chart showing the process of step 88. Update status 11 is "Updating"
If so (step 101 YES), the check for the current record storage address is suspended and step 90
Go to If the update state 11 is "initial" (step 1
01NO), the data record is read from the data section 6 based on the current record storage address, the index key is extracted (step 102), and the index section 5 is searched based on the extracted index key (step 103). . If the index does not exist (NO in step 104), the check result 16 is set to "abnormal" (step 105), and the process proceeds to step 89. If the index exists (step 104 YES) and the record storage address corresponding to the index key matches the record storage address currently being checked (step 106).
YES), the check result 16 is set to "normal" (step 107), and the process goes to step 89. If the index exists (step 104 YES) and the record storage address corresponding to the index key does not match the record storage address currently being checked (step 106N).
O), the check result 16 is set to "abnormal" (step 108), and the process goes to step 89.

The processing of steps 84 and 86 is as follows.
No addition or deletion of the data record occurs between the process of setting the presence / absence of an index for a specific record storage address (step 51, FIG. 6) and the process of setting the data section usage state (step 52, FIG. 7). This is the case in most applications that do not add or delete data records. Step 95
Processing is a case where the data record is deleted between step 51 and step 52 for a specific record storage address, the data part usage state is “unused”, and the corresponding index key is deleted. Therefore, the check result is “normal”. The process of step 98 is a case in which the data record is deleted after the data record is deleted between step 51 and step 52 for the specific record storage address, and the record storage address is moved. , The once deleted index key is added again and remains in the index section 5, but the corresponding record storage address is changed, so the record storage address corresponding to the index key 13 is currently being checked. It was a case that did not match with, and was set as “normal”. However, in a very exceptional case, if there is a possibility that the data record is added after the data record is deleted and the record storage address does not move (YES in step 96), step 98 The result may be set to "abnormal" in the same manner as in step 97, and then the operation may be performed operatively. The process of step 107 is a case where a new data record is added between step 51 and step 52 for a specific record storage address, the data part usage state is “in use”, and the addition of the data record causes Since the presence / absence of the index has been proved, the check result is “normal”. In all other cases, the check results are "abnormal". If the index key cannot be extracted from the data section in step 102, the data record once added may have been further deleted, and the check result may be “abnormal”.

According to this embodiment, the presence / absence of the index 12
To refer to the index part 5 only once in a sequential manner, and to set the data part usage state 14 to refer to the data part 6 only once in a sequential manner. In many cases, the correspondence check table 7 Since the check result "normal" can be obtained only by referring to the index presence / absence 12 and the data part usage state 14, the processing of the correspondence check program 1 can be performed very quickly. Moreover, in exceptional cases, the index section 5 and the data section 6 are accessed again at the stage of the correspondence check (step 53), but at this time, the database management program 3 is updating the information in the database 4. To avoid a temporary mismatch between the index section 5 and the data section 6 during this period.

[0025]

According to the present invention, since the correspondence check program can be run without stopping the business program for updating the database, the serviceability can be improved.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an information processing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration example of an index section and a data section.

FIG. 3 is a diagram showing a data configuration example of a correspondence check table 7.

FIG. 4 is a diagram showing a processing procedure of a database management program for adding a data record.

FIG. 5 is a diagram showing a processing procedure of a correspondence check program of the embodiment.

FIG. 6 is a diagram showing a flow of processing for setting presence / absence of an index according to the embodiment.

FIG. 7 is a diagram showing a flow of processing for setting a data unit use state of the embodiment.

FIG. 8A is a diagram showing a flow of processing of a correspondence check according to the embodiment.

FIG. 8b is an expanded view of step 87 of FIG. 8a.

FIG. 8c is an expanded view of step 88 of FIG. 8a.

[Explanation of symbols]

1 ... Correspondence check program, 2 ... Business program, 3 ... Database management program, 4 ...
Database, 7 ... Correspondence check table, 1
0 ... record storage address, 11 ... update status,
12 ... Presence / absence of index, 13 ... Index key, 14 ... Data section usage state, 15 ... Check target, 16 ... Check result, 17 ... Index key, 18 ... Record Storage address, 22
..Used state, 23 ... Data record

Front Page Continuation (72) Inventor Ninosuke Nakamura 1-6, 1-chome, Uchisaiwai-cho, Chiyoda-ku, Tokyo Nihon Telegraph and Telephone Corporation

Claims (1)

[Claims] 1. An index section and a data section,
The index part stores the correspondence between the index key and the storage address of the data record, the data part stores the database storing the data record, and the storage part corresponds to the storage address when updating the information of the index part and the data part. Processing means for database management that displays a display during updating and whether the index key and data record corresponding to the storage address exist by referring to the database except when the display during updating is displayed An information processing apparatus for checking a database, comprising: a processing unit for checking whether or not it is present.