JPH06139124A - System for deleting unmatch in inverted file system data base - Google Patents

Abstract

PURPOSE:To shorten time required for the restoration of a data base by checking mismatch between data and an index, deleting a unmatched part and leaving only a matched part. CONSTITUTION:This unmatch deleting system consists of a an unmatched part collating processing part 3 for inputting an index part 1 and a data part 2 from an inverted file system data base, mutually collating them and extracting the unmatched part 4 and the unmatched part deleting processing part 5 for inputting the index part 1, the data part 2 and the unmatched part 4 and outputting the unmatch-deleted index part 6 having no unmatch and the unmatch-delected data part 7. Since an index corresponding to the unmatched data record is deleted and only the matched part is left, the original data base can be restored by extracting only the deleted data part, reregistering the extracted part and preparing its corresponding index.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an inverted file method inconsistent deletion method for facilitating failure recovery by extracting and deleting an inconsistent portion between a data section and an index section constituting an inverted file method database. Regarding

[0002]

2. Description of the Related Art An inverted file system database is composed of a data section and an index section. However, if for some reason there is a mismatch between the data section and the index section, system operation such as search processing will be performed. It causes trouble. In the conventional database system, when some kind of system failure occurs and it is recovered, there is no special means to check the inconsistency between the data part and the index part in the failure recovery process, and only the inconsistent part is detected. Since it is impossible to correct it, if there is a possibility of inconsistency, the entire database was recreated from the original data stored on the magnetic tape.

[0003]

In the above-mentioned conventional inverted file system database, when there is a mismatch between the data part and the index part, it is necessary to re-register the original data and recreate the database from the beginning. Therefore, there is a drawback that the time and effort required for re-creation increase as the scale of the database increases.

An object of the present invention is an inverted file system inconsistency deletion method capable of recovering by partial re-creation by detecting and deleting an inconsistency part between a data part and an index part. To provide.

[0005]

An inconsistent deletion method in an inverted file method database according to claim 1 is to input and collate all data of a data section and an index section which constitute the inverted file method database, and inconsistently. An inconsistent part collation processing unit that extracts and outputs a certain data record and index, and an inconsistent part deletion processing unit that deletes the data record and index extracted by this inconsistent part collation processing unit and outputs only the consistent part It is configured with.

In the inverted file method in-database inconsistent deletion method of claim 2, the inconsistent part collation processing section in claim 1 reads the data records of the data section one by one and each data item in the record is read. Is checked against the index part, and when there is a data item for which the corresponding index does not exist, that data record and the corresponding index are extracted as inconsistent parts, and when the matching process is completed for all data records, no matching is performed. When the data record corresponding to the index and the data record corresponding to the index exist in the data part, the data record and the index corresponding thereto are additionally extracted as a mismatched portion.

[0007]

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

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

As shown in FIG. 1, the inverted file method database inconsistency deletion method of this embodiment inputs the index section 1 and the data section 2 from the inverted file method database and collates them to obtain an inconsistent portion. Four
The inconsistent part collation processing unit 3 for extracting the index, the index unit 1, the data unit 2 and the inconsistent part 4 are input, and the inconsistent deleted index unit 6 and the inconsistent deleted data unit 7 having no inconsistency are output. The inconsistent part deletion processing unit 5 is included.

FIG. 2 is a flow chart showing the operation of this embodiment. The specific operation of this embodiment will be described in detail below with reference to FIG.

First, it is assumed that the index section and the data section 2 shown in FIG. 3 are created as the index section 1 and the data section 2 shown in FIG. Here, R1 to R8 shown on the left side of the data part are record numbers, and one record is one data. For example, the first data is managed by the record number R1, and the data items extracted as an index therein include “company”, “internationalization”, “progress”, “foreign country”, and “dispatch”. Further, R1 to R8 in the index section indicate record numbers of data in which the term of the index exists as a data item.

The inconsistent part collation processing unit 3 executes step 11
In step 12, data is acquired one by one from the data part 2, and in step 12, each data item in the acquired data is collated with all the data in the index part 1 that has been previously read and placed, and the index corresponding to each data item is obtained. It is checked whether all are created in the index section 1. If all have been created, step 11
Return to and acquire the next data, and perform the same processing.

The example of FIG. 3 will be specifically described. First, the first data (R1) of the data section is acquired, and for each data item of the data items “company”, “internationalization”, “progress”, “foreign country”, and “dispatch”, there are all indexes corresponding to the index section. Check whether or not. In this case, since all indexes exist, the second data is acquired subsequently. All indexes exist for the second case as well. In the same way, check the third and fourth cases ...
When the inconsistent portion is found in the fifth case, the process is passed to the inconsistent portion deleting processing unit 5 together with the information.

The inconsistent part deletion processing section 5 deletes all data records having an inconsistency in step 13 and all indexes corresponding to the record in step 14 based on the information from the inconsistent part collation processing section 3. .

The data items of the fifth data (R5) in FIG. 3 are "policy", "necessity" and "company", but the indexes created in the inverted file method database are "policy" and "policy". "Necessity" only, "company"
The index corresponding to the first, third and fourth data is not the index corresponding to the fifth data. Therefore, the data is not consistent with the index. In this case, after deleting the fifth data record, the index “policy” and “necessity” corresponding to this are deleted, and the process returns to step 11 to process the sixth data.

The data item of the sixth data item (R6) is "countermeasure" and "foreign work", but the "countermeasure" in the index section is an index corresponding to the fourth data item, and the sixth item is It is not an index corresponding to the data.
Also, there is no index for "foreign work." In this case, since the index corresponding to the sixth data does not exist in the database, only the sixth data record is deleted and the process for the next data is performed.

The data items of the seventh data (R7) are "reward", "opposite" and "foreign". All indexes corresponding to this are created in the index section. Therefore, the seventh data record is not deleted and the eighth data is processed.

As described above, one data record is sequentially acquired from the data part 2 and the inconsistent part with the index part 1 in the inverted file system database is collated. And the index corresponding to the data record are deleted from the inverted file method database until there are no more data records to be acquired.

When there are no more data records to acquire, the process proceeds to step 15 and the unmatched index in the index part 1, that is, the index in which the corresponding data item does not exist (the record number of the index part in FIG. 3 is underlined). Stuff) and delete the data records and indexes associated with it in steps 16 and 17. This eliminates any inconsistencies between the data in the database and the index. The unmatched index can be easily extracted by, for example, deleting the checked index from the read index part, because the index remains when there is no more data to be acquired.

In the example of FIG. 3, since the eighth data does not exist after the processing of the seventh data is completed, the process proceeds to step 15 and the indexes "○" of R7 and R8 underlined as unmatched indexes. ○ "" XX "is extracted. Among them, the data record of the record number corresponding to R7 exists in the data part, but since the data item "XX" is not included in it, there is a mismatch between the data and the index. Also, the data record corresponding to R8 does not exist in the data part. Therefore, in step 16, all the indexes of R7 including "○○" and the indexes "○○" and "XX" of R8 are deleted, and the data record (R7) of the data part corresponding to the index deleted in step 17 is created. delete.

The contents of the inconsistent portion 4 output from the inconsistent portion collation processing portion 3 shown in FIG. 1 are as shown in FIG. 4 in the case of the data portion and the index portion shown in FIG. These pieces of information are passed to the inconsistent portion deletion processing unit 5 to be processed, printed out as inconsistent information, and referred to as data necessary for restoring the database after the inconsistent deletion. Further, the contents of the post-mismatch elimination index part 6 and the post-mismatch elimination data part 7 output from the inconsistency part deletion processing part 5 are as shown in FIG.

In FIG. 2, the inconsistent portion collation processing unit 3 reads the data in the data portion one record at a time and performs collation. If there is an inconsistency, the data record and the corresponding index are immediately discriminated from each other. Deletion processing unit 5
However, even if the data record in which the inconsistency is detected and the index are stored as the inconsistent portion 4 and all the collations are completed and then the inconsistent portion deletion processing unit 5 deletes them collectively. Is the same.

In addition, the inconsistent part collation processing section 3 first searches the index section based on the data record of the data section, extracts the inconsistency in which the corresponding index does not exist, and thereafter, the index in which only the index exists. Although matching is extracted, conversely, it is also possible to search based on the index, first to extract the inconsistency in which only the index exists, and then to extract the data record in which the index does not exist.

[0024]

As described above, according to the inverted file method in-database inconsistency deletion method of the present invention,
If some kind of failure occurs and there is a mismatch between the data part and the index part that make up the inverted file method database, check the consistency between each data and the index, and check the mismatched data record and the corresponding index. Since only the part that is deleted and consistent is output and the information of the deleted data is output at the same time, the original data can be restored by extracting only the deleted data part and re-registering it and creating the corresponding index, This has the effect of simplifying database recovery work.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is a flowchart showing an example of the operation of this embodiment.

FIG. 3 is an explanatory diagram of an example of an index corresponding to a data item including a mismatched portion.

FIG. 4 is an explanatory diagram showing a mismatched portion extracted from FIG.

5 is an explanatory diagram of data items and indexes in which the inconsistent portion is deleted from FIG.

[Explanation of symbols]

 1 Index part 2 Data part 3 Inconsistent part collation processing part 4 Inconsistent part 5 Inconsistent part deletion processing part 6 Index part after inconsistency deletion 7 Data part after inconsistency deletion

Claims (2)

[Claims] 1. An inconsistent part collation processing unit for inputting and collating all data of a data section and an index section constituting an inverted file system database and extracting and outputting a data record and an index having inconsistency, and An inconsistent deletion method in an inverted file system database, which comprises: an inconsistent part deletion processing unit that deletes a data record and an index extracted by the inconsistent part matching processing unit and outputs only a matched part. 2. The inconsistent part collation processing unit reads data records of the data unit one by one and collates each data item in the record with the index unit, and a data item for which a corresponding index does not exist is found. If there is, the data record and its corresponding index are extracted as inconsistent parts, and when the matching process for all data records is completed, the index that remains unmatched and the data record corresponding to that index are stored in the data section. 2. The inverted file type inconsistency deletion method according to claim 1, wherein the data record and the index corresponding thereto are additionally extracted as an inconsistency portion when they exist.