JPH0877044A - Data base controller - Google Patents

Abstract

PURPOSE: To prevent the retrieval process speed of data from decreasing by compressing data stored in a data base in record units, expanding them in a memory, and retrieving the data stored in the memory at a request from an application. CONSTITUTION: The data in the data base stored on a hard disk 2 are compressed in record units and expanded in the memory 3. Then the data stored in the memory 3 are retrieved at the request from the application. Therefore, even when the data base is large in volume, the whole data base can be stored in the memory 3 or even if the whole data base can not be stored in the memory 3, the frequency of swapping can be decreased; and further the compressed data base in the memory 3 can be discriminated, so the retrieval process speed of the data can be prevented from decreasing.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to database management,
The present invention relates to a database control device suitable for use in a database management system for controlling.

[0002]

2. Description of the Related Art FIG. 9 shows a conventional database management system (DBMS).
The structure of an example of the database apparatus to which is applied is shown. A database that stores data in record units is a low-speed but large-capacity hard disk 102.
Is stored as a file. Hard disk 1
The database stored in 02 is read by the CPU 101 and supplied to and stored in a high-speed memory (for example, RAM) 103 having a small capacity (stored in the memory 103).

The CPU 101 executes a predetermined application program, and when there is a request for a data search process from the application program, the CPU 101 searches the database stored in the RAM 103 for the data requested to be searched, and stores it in the application program. It is designed to be handed over.

As described above, in the conventional database apparatus, the data retrieval processing can be performed at high speed by expanding the database in the high-speed memory 103 (compared to the hard disk 102 etc.). There is.

Further, the CPU 101 stores a record that has been accessed and extracted once during one data search process (the record itself may be used, or a code that can identify the record may be used) and then accessed. When the recorded record is the same as the already stored record, the record is not extracted (not processed by the application program).

In this way, in the conventional database device, it is possible to prevent malfunction of the application program due to duplicate processing of data stored in the same record by the application program. ing.

[0007]

By the way, as the data stored in the database increases, the file size thereof also increases. Therefore, if the file size of the database exceeds the capacity of the memory 103, the memory 1
In 03, only part of it can be stored,
As a result, for example, when the entire data of the database is to be searched, it is necessary to replace the data stored in the memory 103, that is, swap (memory swapping).

Here, FIG. 10 shows a state of swapping. Of the databases stored in the hard disk 102, as many records as the memory 103 can store are expanded in the memory 103. Further, for example, a swap area 111 is secured in the hard disk 102 or another secondary storage medium (for example, a RAM disk or the like) (not shown), and records that cannot be stored in the memory 103 are stored therein. Swap area 1
When the record stored in 11 is accessed, of the records stored in the memory 103, for example, a record that has not been accessed recently is transferred to and stored in the swap area 111 and also stored in the swap area 111. The record that was on the memory 10
3 is stored and stored. In this way, memory 1
The data to be stored in 03 is replaced.

When the file size of the database is not so large as compared with the capacity of the memory 103,
Although the number of times of swapping may be small, if the file size of the database becomes considerably larger than the capacity of the memory 103, swapping will frequently occur, and the access speed to the memory 103 will decrease. Therefore, in this case, there is a problem that the data search process becomes slow.

Therefore, as shown in FIG. 11, the database is compressed, its file size is reduced, and the memory 1
Although the method of expanding to 03 is conceivable, if the entire database (entire file) is compressed, the compression result will be different from the original database structure, and the record cannot be accessed as it is. Had to do. For this reason, there is still a problem that the data search process is delayed.

On the other hand, during the above-described one-time data retrieval processing, the record accessed and extracted once is stored, and when the record accessed thereafter is the same as the record already stored, that record is extracted. The processing (processing not to be processed by the application program) is generally performed by the CPU 101 executing a part of the application program. That is, conventionally, the application program is in charge of the above-described processing, and therefore the application program becomes complicated, and a lot of labor is required to create the application program.

Further, in performing the above-mentioned processing, it is necessary to secure an area for storing the record once accessed and extracted, for example, in the memory 103. In this case,
In the worst case, the number of records stored in the memory 103 will be the same as the number of records in the database, so the storage area of the memory 103 will be compressed, and this will cause frequent swapping. There was a problem that the search process of was slowed down.

The present invention has been made in view of such a situation, and it is possible to prevent a decrease in data retrieval processing speed.

[0014]

A first database control device of the present invention is a database storing data and a memory for expanding the data stored in the database (for example, the memory 3 shown in FIG. 1). A database control device for controlling a database device comprising:
The data stored in the database is compressed in record units, expanded in a memory, and the data stored in the memory is searched in response to a request from an application.

In this database control device, the data stored in the database can be compressed in record units only when swapping is required to expand the data stored in the database into the memory.

A second database control device of the present invention is
A database control device that retrieves data of a predetermined record from data of a database in response to a request from an application, and the record is a value storage unit (for example, a flag storage unit illustrated in FIG. 3) that stores a predetermined value. Such)
And a data storage unit for storing data (for example, the data storage unit shown in FIG. 3), a unique value is assigned to a request from an application, and a search is performed according to the request from the application. When the value stored in the value storage of the record in which the data is stored differs from the unique value assigned to the request from the application, the data stored in the data storage of the record is passed to the application, It is characterized in that a unique value is stored in the value storage unit of the record in which the data retrieved in response to the request is stored.

In this database control device, the unique value can be used as the time stamp corresponding to the date and time when the application makes a request. Further, when the record belongs to a parent of a predetermined attribute, in response to a request from the application, a condition for performing a data search includes a parent limiting condition that limits the parent to which the data belongs,
A unique value can be stored in the value storage unit of a record that contains at least the parent limitation condition.

[0018]

In the first database control apparatus of the present invention, the data of the database stored in the hard disk 2, for example, is compressed in record units and expanded in the memory 3. Then, the data stored in the memory 3 is searched in response to the request from the application. Therefore, even if the database has a large capacity, the entire database can be stored in the memory 3, or even if the entire database cannot be stored in the memory 3, the number of swapping can be reduced. Further, since it is possible to identify the record of the compressed database on the memory 3, it is possible to prevent the reduction of the data search processing speed.

In the second database control apparatus of the present invention, the data of the predetermined record is retrieved from the data of the database in response to the request from the application. In this case, a value unique to the request is assigned to the request from the application, and the storage value of the flag storage unit of the record in which the data retrieved in response to the request from the application is stored corresponds to the request from the application. When it is different from the assigned unique value,
The data stored in the data storage unit of the record is passed to the application. Then, a unique value is stored in the flag storage unit of the record in which the data retrieved in response to the request from the application is stored. Therefore, in order to prevent re-extraction of a record that has been accessed and extracted once during one data retrieval process, it is not necessary to store the record, and the storage area of the memory of the device is squeezed. Can be prevented. As a result, it is possible to prevent a decrease in the data search processing speed.

[0020]

FIG. 1 shows the configuration of an embodiment of a database device managed and controlled by a database management system to which the present invention is applied. The database management system is realized by, for example, software, and the program is stored (recorded) in the hard disk 2, which is a medium having a large capacity at a low speed.
Have been.

The hard disk 2 stores a database, an application program that uses the database, and the like, in addition to the programs of the database management system described above. The CPU 1 executes a program stored in the hard disk 2 as a medium having a small capacity but high speed, for example, a memory (for example, R
It is designed to be loaded and executed in 3). Further, the CPU 1 is adapted to perform a predetermined process according to the operation of the pointing device 5 or the keyboard 6. Further, the CPU 1 is also configured to perform display control of the CRT 4 and read / write control of data with respect to the hard disk 2 and the memory 3.

The memory 3 stores data and programs necessary for the operation of the CPU 1. C
The RT4 is adapted to display according to the control of the CPU1. The pointing device 5 and the keyboard 6 are operated when performing a predetermined input to the CPU 1. The pointing device 5 is composed of, for example, a mouse or a pen.

Next, the operation will be described. When the device is started up, the program of the database management system stored in the hard disk 2 is loaded into the memory 3. Then, the CPU 1 executes the program of the database management system loaded in the memory 3 to expand the database in the memory 3. That is, as shown in FIG. 2, the CPU 1 compresses the data of the database stored in the hard disk 2 in record units, transfers the data to the memory 3, and stores the compressed data (decompresses in the memory 3).

After that, when the pointing device 5 and the keyboard 6 are operated to instruct the CPU 1 to start the application program stored in the hard disk 2, the CPU 1 loads the application program into the memory 3. To execute.

When a database search request is issued from the application program, the CPU 1
The data corresponding to the request is searched from the database data stored in the memory 3 according to the program of the database management system.

Since the database stored in the memory 3 is compressed in record units as described above, the record can be identified. Therefore, C
PU1 can retrieve the data according to the request of the application program from the compressed database.

When the CPU 1 obtains the data according to the request of the application program, the CPU 1 decompresses the data according to the program of the database management system and transfers the data to the application program. In the application program, a predetermined process is performed using the data obtained as a result of the search (this process is performed by the CPU 1
Is done according to the application program).

As described above, the database is compressed so that records can be identified. Therefore, even if the database has a large capacity, the number of swapping operations can be reduced and the data retrieval process can be performed. It is possible to prevent the speed from decreasing. Further, if the databases have the same capacity, the data retrieval processing speed can be improved as compared with the conventional case.

Further, data compression and decompression processing are
Since it is performed by the program of the database management system, it is not necessary to include such processing in the application program, so that the application program can be easily created.

It should be noted that the database can be compressed only when the database has a larger capacity than the storage capacity of the memory 3 and therefore the database is directly expanded in the memory 3 when swapping is required. . This can be performed by causing the program of the database management system to compare the capacity of the database with the storage capacity of the memory 3.

Next, in the database device of FIG.
Each record of the database stored in the hard disk 2 is composed of a record number storage unit, a flag storage unit, and a data storage unit, as shown in FIG. 3, for example. The record number storage unit stores a record number Rm as a unique identification number assigned to the record. Flag F is stored in the flag storage unit.
(Details will be described later) is stored. Actual data is stored in the data storage unit.

The records configured as described above are classified based on, for example, the data stored in the data storage unit, and according to the classification result, as shown in FIG. belong to. Note that a record does not belong to only one parent, for example, as shown in FIG.
It may belong to a plurality of parents (parents Pn and Pn + 1 in FIG. 5).

Here, for example, as shown in FIG.
Consider a database consisting of five records with record numbers R1 to R5. In the figure, an arrow having each parent P1 to P6 as a starting point and each record as an ending point means that the record at the ending point belongs to the parent P at the starting point.

For a database as shown in FIG. 6, for example, a search condition "a record belongs to a parent P2n (a parent whose number following P is an even number) and the data stored in the data storage section is 50 or more", When there is a request for data search processing from an application program and a search is performed without any particular restriction, the following search results are passed to the application program.

That is, in this case, first, the records of the record numbers R1 and R5 belonging to the parent P2 are set as the search result candidates. (However, the search is performed from left to right in the figure).

The data stored in the data storage section of the record of record number R1 or R5 is 4 respectively.
2 or 72, and in this case, therefore, the record of the record number R5 having 50 or more data matching the search condition is extracted.

Then, next, the record of the record number R3 belonging to the parent P4 is made a candidate for the search result. The data stored in the data storage section of this record is 55. Therefore, in this case, the record with the record number R3 matches the search condition and is extracted as the search result.

Next, the records of record numbers R3 and R5 belonging to the parent P6 are candidates for the search result. The data stored in the data storage unit of these records all match the search condition, and therefore, in this case, both the records of record numbers R3 and R5 are extracted again as the search result.

From the above, the records with record numbers R5, R3, R3, and R5 are sequentially obtained as the search result, and the record with the record number R3 or R5 is obtained.
Each is duplicated and passed to the application program.

In this case, in the application program, the data stored in the records of record numbers R3 and R5 will be processed in duplicate, and therefore the same record will be passed in duplicate to the application program. Need to be prevented.

Therefore, in the database apparatus of FIG. 1, when the data of a predetermined record is searched from the data of the database in response to the request from the application program, first, in response to the request from the application program, A unique value is assigned to the request (this is done by the CPU 1 according to the program of the database management system).

Here, as the unique value assigned to the request from the application program, for example, a time stamp corresponding to the date and time of the request is used.

When the CPU 1 performs the search as described above and a record matching the search condition is obtained, the flag F stored in the flag storage section of the record is assigned to the request from the application. It is determined whether the time stamps match. And
The flag F stored in the flag storage unit of the record is
If it is determined that the record matches the time stamp, the record is not passed to the application program, and only if it is determined to be different (does not match), the record is passed to the application program.

After that, the time stamp is stored as the flag F in the flag storage portion of the record in the database (the time stamp is reflected in the flag F).

From the above, in the database apparatus of FIG. 1, when the application program performs the data retrieval processing under the retrieval condition similar to that in FIG. 6, it is performed as follows.

That is, first, for a request from an application program, a time stamp is assigned to the request. Note that the time stamp is set to 1, for example, in order to simplify the description. The flag F recorded in the flag storage unit of each record of the database is a value other than 1 as shown in FIG.
For example, it is assumed to be 0.

In this case, as described above, first, the record of R5 of the record numbers R1 and R5 belonging to the parent P2 is extracted. Then, it is determined whether or not the flag F stored in the flag storage unit of this record matches the time stamp assigned to the request from the application. In this case, since the flag F of the record of the record number R5 is 0, it is determined that the flag F is different from the time stamp, and this record is passed to the application program.

Then, as shown in FIG. 7, the time stamp is stored as the flag F in the flag storage portion of the record of the record number R5. That is, the flag F of the record of the record number R5 is set to 1.

Next, the record of the record number R3 belonging to the parent P4 is extracted. Now, the flag F of this record
Is 0, the record with the record number R3 is passed to the application program, like the record with the record number R5 described above, and further, as shown in FIG. It is stored as a flag F.

Further, next, records with record numbers R3 and R5 belonging to the parent P6 are extracted as the retrieval result. However, in this case, the flags F of the records of the record numbers R3 and R5 are both set to 1 as described above, and therefore, it is determined that any of the flags F matches the time stamp. Therefore, in this case, neither of the records of record numbers R3 and R5 is passed to the application program.

From the above, the application programs are provided with record numbers R3 and R5 that match the search conditions.
Will be passed one by one.

Therefore, during one data retrieval process (retrieval process performed according to an application program, for example, transaction process), the record once accessed and extracted again is extracted (passed to the application program). ), It is not necessary to store the record, so that it is possible to prevent the storage area of the memory 3 from being pressed. As a result, it is possible to prevent a decrease in the data search processing speed.

In the above-mentioned case, after the record matching the search condition is obtained, the time stamp is stored as the flag F in the flag storage section of the record. When a record that matches the subscript m of the parent Pm is an even number in the above case is obtained, a time stamp may be set in the flag storage unit of the record.

That is, for example, in the case described with reference to FIG. 7, as shown in FIG. 8, all the search conditions are not met, but the conditions regarding the parent among them are met, for example, of the record number R1 belonging to the parent P2. A time stamp can be set as the flag F in the flag storage unit of the record.

In this case, when the record of record number R1 belongs to not only the parent P2 but also the parents P4 and P6, this record is not only searched for the parent P2 but also for the parents P4 and P6. Although it is a candidate for the search result, since the flag F matches the time stamp, it can be excluded from the search result without evaluating the data stored in the data storage unit. Therefore, in this case, the speed of the search process can be improved.

Further, the search as described with reference to FIGS. 3 to 8 can be performed by expanding the database in the memory 3 or can be performed on the database stored in the hard disk 2.

Further, in the case described with reference to FIGS. 3 to 8, the database management system determines all the search conditions. However, for example, the database management system is caused to determine only the above-mentioned conditions relating to the parent. The other conditions for the data stored in the data storage unit can be determined by the application program (the application program is evaluated).

In the case described with reference to FIGS. 3 to 8,
Although the time stamp is set in the flag storage unit, in addition to this, for example, a predetermined bit in the flag storage unit of the once-extracted record is set, and the extracted record is applied to the application based on the bit. It is also possible to decide whether to pass it to the program. However, in this case, since it is necessary to clear the bit each time the search processing is completed, and therefore the processing becomes complicated, as described above, a value unique to the request of the application program such as a time stamp is used. It is preferable to do so. That is, when the unique value assigned to the request of the application program is used, it is not necessary to clear the flag storage unit, and it is possible to prevent complication of processing and complication of the program of the database management system.

Further, the unique value assigned to the request of the application program is not limited to the time stamp.

In this embodiment, the data is searched in response to the request from the application program. However, not only the request from the software program but also the request from the terminal or other hardware is performed. It is also possible to search for data (in this specification, these are collectively referred to as an application).

[0061]

As described above, according to the database control apparatus of the present invention, it is possible to prevent a decrease in the data retrieval processing speed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a database device to which the present invention is applied.

FIG. 2 is a diagram showing how a database stored in a hard disk 2 is compressed and expanded in a memory 3 in the embodiment shown in FIG.

3 is a diagram showing a structure of a record of a database stored in a hard disk 2 of FIG.

FIG. 4 is a diagram showing that a record belongs to a parent P.

FIG. 5 is a diagram showing that a record belongs to parents Pn and Pn + 1.

6 is a diagram showing a database stored in a hard disk 2 of FIG.

FIG. 7 is a diagram for explaining a database search process performed in the embodiment of FIG.

8 is a diagram for explaining a database search process performed in the embodiment of FIG. 1. FIG.

FIG. 9 is a block diagram showing a configuration of an example of a conventional database device.

FIG. 10 is a diagram for explaining swapping.

FIG. 11 is a diagram illustrating a conventional database compression method.

[Explanation of symbols]

 1 CPU 2 Hard Disk 3 Memory 4 CRT 5 Pointing Device 6 Keyboard 101 CPU 102 Hard Disk 103 Memory

Claims (5)

[Claims] 1. A database control device for controlling a database device comprising: a database storing data; and a memory for expanding the data stored in the database, wherein the data stored in the database is recorded as a record. A database control device, which compresses a unit, expands it in the memory, and retrieves data stored in the memory in response to a request from an application. 2. The data stored in the database is compressed in record units only when swapping is required to expand the data stored in the database into the memory. The database device described. 3. In response to a request from the application,
A database control device that retrieves data of a predetermined record from data of a database, wherein the record includes a value storage unit that stores a predetermined value and a data storage unit that stores the data. The request is assigned a unique value to the request, and the storage value of the value storage unit of the record storing the data retrieved in response to the request from the application is assigned to the request from the application. When the value is different from the unique value, the data stored in the data storage unit of the record is passed to the application, and the value storage unit of the record in which the data retrieved in response to a request from the application is stored The database controller stores the unique value in the. 4. The database control device according to claim 3, wherein the unique value is a time stamp corresponding to a date and time when a request is made from the application. 5. The record belongs to a parent having a predetermined attribute, and the condition when performing a search for the data in response to a request from the application is a parent limitation that limits the parent to which the data belongs. The database control device according to claim 3, wherein the unique value is stored in the value storage unit of a record that includes at least a condition and that matches a parental limit condition.