JPH04330534A - File retrieval system - Google Patents

Abstract

PURPOSE:To increase the speed of record insertion or change processing independently of the number of indexes set to a record. CONSTITUTION:When a record is inserted to or changed in a data file storage part 14, index information corresponding to the new record as the insertion or change object is only stored in a memory 12 but is not written in an indexed file storage part 13. Index information stored in the memory 12 is written in the indexed file storage part 13 at the time of system stop, and thereby, index information in the indexed file storage part 13 is reconstituted. Thus, the record insertion or change processing is executed at a high speed independently of the number of indexes set to the record.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention This invention relates to a file search system, and more particularly to a file search system that uses an index to search for file data.

0002

2. Description of the Related Art In recent years, various file search systems have been developed that search for file data using indexes. This file search system is configured to efficiently retrieve records from files by preparing an index.

[0003] In other words, in a file search system using this index, when a record is written to a data file, an index is created that indicates the correspondence between data items in the record and storage positions on the data file. will be registered. When reading a record from a data file, the storage position of the record is checked using the index on the index file, and the record at that storage position is read.

[0004] In this way, by performing a record search using an index, by knowing the storage position of the record, the record can be read directly from the data file, and the record search efficiency can be improved.

[0005] Recently, in order to further improve record search efficiency, indexes are increasingly being provided not only for one data item but also for multiple types of data items included in a record.

For example, in a database file used in a company's human resources management system, among the data items included in employee records in the file,
Indexes are set for each data item indicating the employee number, the date of joining, the employee name, and the salary. In this way, it becomes possible to find the necessary record using any of the employee number, date of joining, employee name, and salary, and the search can be performed efficiently.

[0007] However, when multiple types of data items for indexes are set in this way, the number of types of index information to be created on the index file also increases, so when writing records to the data file, Therefore, it becomes necessary to write index information for each type (number) of indexes.

[0008] For this reason, for example, as mentioned above, the record contains the employee number, date of joining, employee name, and salary.
If a type index is set, each time a record is written to the data file, it is necessary to write to the index file four times.

[0009] Therefore, the total number of times of writing required to register one employee record in the data file is five times, including writing the employee record to the data file (one time).

As described above, setting a plurality of types of indexes can improve record retrieval efficiency, but at the same time increases the number of writes to the index file. Therefore, if multiple types of indexes are set, data must be written to the index file many times each time the data file is updated (for example, by inserting or changing a record). This may cause problems such as the change process taking a long time.

[0011]

Conventionally, setting a plurality of types of indexes can improve record retrieval efficiency, but has the disadvantage that record insertion and modification processing takes time.

The present invention has been made in view of the above points, and provides a file search system that can perform record insertion and modification processing at high speed regardless of the number of indexes set for a record. The purpose is to

[0013]

[Means and operations for solving the problems] A file search system according to the present invention includes a first file that stores file data consisting of a plurality of records, and storage positions of the plurality of records on the first file. , a second file that stores index information shown in association with predetermined data items included in the records;
a search means for searching file data on a file; a data update means for inserting or changing a record in the first file; and a data updating means for inserting or changing a record in the first file; means for creating corresponding index information and storing it on a main memory; and when the system stops, writing the index information stored on the main memory into the second file to reconstruct the index information in the second file. It is characterized by comprising means.

[0014] In this file search system, when a record is inserted or changed, the index information corresponding to the new record to be inserted or changed is only stored in the memory, and the index information is not stored in the second file. Not written. The index information stored in memory is written to a second file when the system is stopped, thereby
File index information is reorganized. therefore,
Even if multiple types of indexes are set for a record,
When inserting or changing a record, the record is only written to the first file and not written to the second file, so inserting or changing records can be performed quickly regardless of the number of indexes set for the record. I can do that.

[0015]

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows the system configuration of a file search system according to an embodiment of the present invention. This file search system is realized on a computer system equipped with a CPU, a main memory, a magnetic disk device, etc., and includes a control unit 11, a memory 12, an index file storage unit 13, a data file storage unit 14, and It is composed of a battery backup mechanism 15.

The control unit 11 is constituted by a CPU, and executes data file search processing and record update processing (change, insertion, deletion) in the data file according to a predetermined program. When performing record modification or insertion processing, the control unit 11 generates index information corresponding to the record to be modified or inserted, and stores it in the memory 12. Further, during the search process, the control unit 11 first checks the index information in the memory 12, and if there is no index corresponding to this index information, checks the index information in the index file storage unit 13. The memory 12 is a storage area allocated to a part of the main memory, and temporarily holds index information corresponding to a new record to be changed or inserted.

The index file storage unit 13 is composed of a magnetic disk device, and stores index information for searching data files in the data file storage unit 14. This index information indicates the correspondence between data items in records constituting the data file and storage positions on the data file, and is provided for each type of index set in the record. A specific example of this index information will be described later with reference to FIG.

The data file storage section 14 is composed of a magnetic disk device, and stores a data file consisting of a plurality of records. Each record is
Contains multiple data items (search keys) used as an index.

The battery backup mechanism 15 uses a battery to transfer the index information stored in the memory 12 to the index file storage unit 1 when the system is stopped such as when the power is turned off.
Write in 3.

FIG. 2 shows an example of specific contents of the data file in the data file storage section 14 and the index file in the index file storage section 13. Here, each record consists of multiple data items indicating employee number, employee name, date of joining, etc. Among these data items, two data items are employee number and employee name. The case where is used as an index will be explained.

[0022] When the two data items of the employee number and the employee name are indexed in this way, the index file in the index file storage section 13 has two types, the first and second types, as shown in the figure. Index information 131 and 132 are set.

The first index information 131 indicates the storage position of a record in association with the data item of employee name, and holds a pointer indicating the storage position of the record for each employee name. In the example in Figure 2, the employee name “Ando”
A pointer E1 is assigned to ``Kikuchi'', a pointer Ei is assigned to the employee name "Kikuchi", and a pointer En is assigned to the employee name "Watanabe", and these pointers E1, Ei, and En are assigned as follows.
The storage positions of records R1, Ri, and Rn on the data file are shown. This storage location may be either a physical storage location on the data file storage unit 14 or a logical storage location such as a record number.

The second index information 132 indicates the storage position of the record in association with the data item of the employee number, and holds a pointer indicating the storage position of the record for each employee number. In the example of Figure 2, the employee number “
Pointer E1 for "0001", employee number "00"
0i”, pointer Ei, employee number “000n”
A pointer En is assigned to , and these pointers E1, Ei, and En indicate the storage positions of records R1, Ri, and Rn, respectively, on the data file. This storage location may be either a physical storage location on the data file storage unit 14 or a logical storage location such as a record number. Here, the record R shown in FIG.
Consider the case where x is inserted into a data file in the data file storage section 14.

[0025] When writing the record Rx to the data file, usually the index information 131, 132 of the index file
Writing to the index file storage unit 13 is also performed at the same time in order to reconstruct the record Rx.
The index information is only written to the memory 12 and not to the index file storage section 13.

In this case, the contents of the index information written to the memory 12 are as shown in FIG. That is, memory 12
contains index information indicating the correspondence between the employee number “000Q” of the record Rx and the pointer Ek indicating the storage location of the record Rx, and the employee name “Sato” of the record Rx and the storage location of the record Rx. Index information indicating the correspondence with the pointer Ek is stored. Next, consider the case where the employee name of record Ri in the data file is changed from "Kikuchi" to "Sato".

In this case, the record Ri is deleted from the data file storage section 14, and the record Ri' shown in FIG. 5 is written into the data file storage section 14 instead. At this time, as shown in FIG. 6, the memory 12 contains information indicating the correspondence between the employee name "Kikuchi" of the old record Ri and the pointer Ei indicating the storage position of the record Ri, and information indicating the correspondence between the employee name "Kikuchi" of the old record Ri and the employee name of the new record Ri'. Information indicating the correspondence between the member name "Sato" and the pointer Ei indicating the storage position of the record Ri' is stored as one piece of index information.

In this way, the reason why the information about the new and old records is held as index information is that when the index information on the memory 12 is reflected in the index file, the employee name "
This is to register "Sato" in place of "Kikuchi" in the index information 131.Next, the operation of the file search system of FIG. 1 will be explained with reference to the flowcharts of FIGS. 7 to 10. First, FIG. The operation when inserting a record into a data file is explained with reference to the flowchart.

In this insertion process, the control unit 11:
First, it is determined whether an index is defined for the record (step S11), and if an index is defined, the index information (see FIG. 4) corresponding to the new record to be inserted (record Rx in FIG. 3) is determined. is created and written into the memory 12 (step S12). The process of step S12 is repeated for the number of indexes that have been set.

When the control unit 11 finishes registering index information for all indexes in the memory 12 (step A13)
, the new record to be inserted is written in the data file storage unit 14 (step S14), and the insertion process is thus completed. Next, with reference to the flowchart of FIG. 8, the operation when changing a record in a data file will be described.

In this change process, the control section 11
First, it is determined whether the data item used as an index is subject to change (step S21), and if it is subject to change, the information of the old record before change and the new record after change are determined. Index information (see Figure 6)
and write it into the memory 12 (step S2
2). The process of step S22 is repeated for the number of indexes that have been set.

[0032] When the control unit 11 finishes registering index information for all indexes in the memory 12 (step A23),
, a new record is written in the data file storage unit 14 at the storage location of the old record (step 24), and the change process is thus completed. Next, with reference to the flowchart of FIG. 9, the operation of search processing for data files will be described.

[0033] In this search process, the control unit 11
First, the index information in the memory 12 is checked to detect whether there is an index corresponding to the search key in the index information (step S31). If there is a corresponding index in the memory 12, the control unit 11 uses the pointer defined in the index information of the memory 12 to access the data file storage unit 14.
and read the corresponding record (step S
33).

On the other hand, if there is no corresponding index in the memory 12, the control unit 11 reads the index information from the index file storage unit 13, and uses the pointer defined in the index information to store the index in the data file storage unit 14. access and
The corresponding record is read (step S32). next,
The operation when the system is stopped will be described with reference to the flowchart in FIG.

When the system is stopped, such as when the power is turned off, the control unit 11 first deletes the information of the old record stored in the memory 12 (part of the index information stored when the record was changed). (Step S41) Next, the index information on the memory 12 is written to the index file storage section 13, and the index information is reconfigured (Step S42).

As described above, in the file search system of this embodiment, when inserting or changing a record into the data file storage unit 14, the index information corresponding to the new record to be inserted or changed is is only accumulated on the memory 12 and is not written to the index file storage section 13. The index information stored in the memory 12 is written to the index file storage section 13 when the system is stopped, and thereby the index information in the index file storage section 13 is reconfigured.

Therefore, even if multiple types of indexes are set for a record, when a record is inserted or changed, the record is simply written to the data file storage section 14, and the index information is not written to the index file storage section 13. is not executed, so record insertion or modification processing can be executed at high speed regardless of the number of indexes set for the record.

[0038]

As described above, according to the present invention, record insertion and modification processing can be executed at high speed regardless of the number of indexes set for a record.

[Brief explanation of drawings]

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

FIG. 2 is a diagram showing an example of a specific configuration of a data file and an index file in the same embodiment.

FIG. 3 is a diagram showing an example of the structure of records inserted into the data file of the same embodiment.

FIG. 4 is a diagram showing an example of index information stored in a memory provided in the embodiment.

FIG. 5 is a diagram showing an example of the structure of a new record used when changing the data file of the same embodiment.

FIG. 6 is a diagram showing another example of index information stored in the memory provided in the embodiment.

FIG. 7 is a flowchart illustrating the operation of record insertion processing in the same embodiment.

FIG. 8 is a flowchart illustrating the operation of record change processing in the same embodiment.

FIG. 9 is a flowchart illustrating the operation of search processing in the same embodiment.

FIG. 10 is a flowchart illustrating the operation when the system is stopped in the same embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 11... Control part, 12... Memory, 13... Index file storage part, 14... Data file storage part, 15... Battery backup mechanism.

Claims (2)

[Claims] 1. A first file that stores file data consisting of a plurality of records, and storage positions of the plurality of records on the first file are associated with predetermined data items included in those records. a second file for storing index information indicating index information; a search means for searching file data on the first file; a data update means for inserting or changing a record in the first file; means for creating and storing index information corresponding to the new record to be inserted or changed each time the system is executed, and storing the index information stored in the memory in the second file when the system is stopped; A file search system comprising: means for reconstructing index information of the second file. 2. The search means refers to index information in the memory and the second file in order, and uses the index information on the main memory and the second file to search for a record to be searched on the first file. 2. The file search system according to claim 1, wherein the storage location of the file search system is checked.