JPH04256139A - Data base control system - Google Patents

Abstract

PURPOSE:To attain a date base control system that can reduce its scale. CONSTITUTION:A relational date base control system is provided with a cursor setting means 121 which analyzes the retrieving conditions end then produces end stores the retrieving result information consisting of the arrangement of the information which identifies the rows of the row numbers, etc., of a real data table 200 obtained from execution of the retrieval based on the analyzing result of the retrieving conditions, and a row retrieving means 122 which retrieves a specific row of the reel date table 200 that satisfies the retrieving conditions by reference to the retrieving result information. The retrieving result information has the same structure as the real, date row end stored in a retrieving result information table 230, i.e., a DB table similar to the table 200. Then the retrieving result information can be used.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a relational database management system (RDBMS), and more particularly to a database management method that can reduce the system scale.

0002

[Prior Art] As a prior art related to a language used for database management, for example, "JIS
3005-1987 [ISO 9075-1987
(E)] Database Language SQL" (Japanese Standards Association, November 1986), etc., is known.

[0003] Database language SQ according to this prior art
When the cursor is opened, L creates search result information in which rows that satisfy the search conditions in the actual data table are sequentially arranged according to the specification. In this prior art, data operations for positioning a cursor and searching, changing, or deleting the positioned row are defined for the created search result information.

[0004] Also, database management systems (DBM)
As a conventional technology related to S), for example, “GIGA
The technique described in "Database Administrator's Guide" (Japan Business Consultant, November 1980, pp. 4-7) is known.

[0005] In this prior art, search result information created when a cursor is opened is a record (row) ID list (array) of an actual data table to be searched. Furthermore, this search result information is too large to fit in the memory and needs to be stored in an external storage device such as a disk, so it is managed by being stored in a dedicated work storage (file).

FIG. 11 is a block diagram showing the system configuration of a relational database management system according to the prior art. In FIG. 11, 600 is an application (AP)
Process group, 610, relational database system, 6
11 is a transaction control unit, 620 is a data operation unit, 630 is an input/output control unit, 640 is a search result information file, and 650 is a relational database (RDB).

In the illustrated prior art, a relational database management system (RDBMS) 610 includes a transaction control section 611, a data manipulation section 620, and an input/output control section 630. The transaction control unit 611 controls the application (AP) process group 6
Execution control of each transaction control command requested by 00 and data consistency control of each data manipulation command are performed.

[0008] The data operation unit 620 also analyzes the search conditions to generate search result information consisting of a sequence of row numbers of the actual data table that satisfies the conditions, and controls the storage of the search result information, a cursor setting control unit 621, and a search function. A row search control unit 622 that controls the row search operation of the actual data table by referring to the result information, a row insertion control unit 623 that controls the row insertion operation, a row update control unit 624 that controls the row modification operation, It also includes a row deletion control unit 625 that controls row deletion operations.

Furthermore, the input/output control unit 630 creates a search result information file 640 that stores the created search result information.
A record input/output processing unit 631 performs input/output processing of the data, and an input/output processing unit 632 performs input/output processing of rows of a relational database (RDB) 650 including a plurality of real data tables 651.

[0010] The search result information is sent to the AP process 601.
When processing a request to retrieve a real data table from the RDBMS
This is internal information that is only referenced by others. Therefore, when manipulating search result information, unlike actual data manipulated by multiple AP processes, a transaction control function for guaranteeing data consistency is not required. Further, the search result information is variable length data consisting of a sequence of information identifying rows with a fixed length of several bytes, and is not compatible with a database table consisting of rows and columns. Therefore, even though it is stored in an external storage device like real data,
It was managed using a dedicated file format and file management system.

[0011]

[Problems to be Solved by the Invention] The above-mentioned prior art uses an actual data table of a relational database (RDB) to store actual data, but it is difficult to identify rows such as row numbers that are search result information. The list of information to be searched is stored in a dedicated file (search result information file) and managed by a dedicated file management system.

[0012] For this reason, the above-mentioned prior art requires a dedicated means for managing the search result information file within the database management system, which has the problem of increasing the scale of the database management system.

An object of the present invention is to solve the problems of the prior art, eliminate the need for dedicated means for file management, and
An object of the present invention is to provide a database management method that can reduce the scale of a database management system.

[0014]

[Means for Solving the Problems] According to the present invention, the object is to analyze a search condition and, based on the result of the analysis, identify a row such as a row number of an actual data table obtained by executing a search. a relational type comprising: a cursor setting means for generating and storing search result information consisting of a sequence of information; and a row search means for referring to the search result information and searching for a specific row of an actual data table that satisfies the search condition. In the database management system, means for storing the search result information in a DB table having the same structure as the actual data row and having the same structure as the actual data table;
This is achieved by comprising means for reading the search result information from a DB table in which the information is stored.

[0015]

[Operation] Since the search result information is stored in the DB table in the same way as the actual data, input/output processing for storing and referencing the search result information can be performed by the same means as for the actual data. Therefore, a special means for inputting and outputting search result information is not required, and the scale of the relational database management system can be reduced.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the database management system according to the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the system configuration of an embodiment of the present invention. 100 is an application (A
P) Process group, 110 is a relational database system, 111 is a transaction control unit, 120 is a data manipulation unit, 130 is an input/output control unit, and 140 is a relational database (RDB).

In one embodiment of the present invention, a relational database management system (RDBMS) 110 includes a transaction control section 111, a data manipulation section 120, and an input/output control section 130. The transaction control unit 111 performs execution control of data operations in each transaction requested by the application (AP) process group 100 and data consistency control of each data operation command.

Furthermore, the data manipulation unit 120 analyzes the search conditions to generate search result information consisting of a row number sequence of the actual data table that satisfies the conditions, and a cursor setting control unit 121 that controls the storage of this information. , refer to the search result information,
Row search control unit 12 that controls row search operations in the actual data table
2. A row insertion control section 123 that controls a row insertion operation, a row update control section 124 that controls a row modification operation, and a row deletion control section 125 that controls a row deletion operation. Furthermore, the input/output control unit 130 uses a relational database (RDB) consisting of a search result information table 230 that stores created search result information and a plurality of actual data tables 200.
It consists only of a line input/output processing section 131 that performs input/output processing of 140 lines.

FIGS. 2, 3 and 4 show actual data tables 200,
3 is a diagram illustrating the data structure of a search result information table 230 and a cursor management information table 270 that manages the inside of memory. FIG.

The real data table 200 shown in FIG. 2 includes table management information 201 consisting of information indicating the logical and physical structure of the real data table 200, and one or more real data storage blocks 21.
It consists of 0. In addition, the actual data storage block 21
0 is block management information 211 consisting of the length of unused area 212 in the block, and one or more actual data rows 2
It consists of 20. Furthermore, the actual data row 220 is
It is composed of row management information 221 including the length of the actual data row 220, and one or more actual data (column values) 222.

The search result information table 230 shown in FIG. 3 is composed of table management information 231 consisting of information indicating the logical and physical structure of the search result information table 230, and one or more search result information storage blocks 240. Ru. The search result information storage block 240 also has an unused area 24 within the block.
The block management information 241 has a length of 2, etc., and one or more search result information rows 250. moreover,
The search result information line 250 includes line management information 251 including the length of the search result information line 250, a line number 252 of the next search result information line, and a line number list consisting of a sequence of line numbers that satisfy the search conditions. It consists of a sublist 260. The line number sublist 260 also includes one or more line numbers 2
It consists of 61 arrays.

The actual data table 200 and the search result information table 230 described above have the same structure.

Cursor management information table 27 shown in FIG.
0 is the cursor name 271, which is set for each cursor setting.
, the row number 272 of the first row among the plurality of search result information rows 250 stored in the search result information table 230, the total number of stored row numbers 273 stored in the search result information table 230, and the currently positioned actual data row. The current line number position 274 indicates the position of the line number 220 in the line number list.

FIG. 5 is a diagram showing an actual data table as an example of a search processing target, FIG. 6 is a diagram explaining a cursor setting request, and FIG.
is a diagram showing the relationship between the line number list obtained by the conditional term and the line number list obtained by combining with the operator, FIG. 8 is a flowchart showing the processing of the cursor setting control unit 121, and FIGS. 9 and 10 are This is a flowchart showing the processing of the row search control unit 122. Next, the operation of an embodiment of the present invention will be described with reference to these figures.

First, the operation of the cursor setting control section 121 will be explained using the flowchart shown in FIG.

(1) Actual data table 300 as shown in FIG.
Assume that an SQL command (cursor setting request) 310 as shown in FIG. 6 is issued in order to set a cursor on a cursor. At the start of processing, the cursor setting control unit 121 stores a search condition consisting of a cursor name 321, a table name 322 of the table 300 shown in FIG. A cursor open request 330 for which the cursor declaration 320 has been completed is passed (step 400).

(2) This passed cursor setting request 31
0, a search condition consisting of condition terms 323 and 324 and a logical operator 325 that connects the condition terms is analyzed, and a list of line numbers as shown in FIG. Create 352. Here, the logical operator AND 325 matches and connects the line number lists 350 and 351 obtained in each condition term 323 and 324 to create a line number list 352 that satisfies the search condition (
step 410).

(3) From the total number m of line numbers in the line number list obtained in step 410, the number n of search result information lines is determined using the maximum number k that can be stored in the line number sublist 260. The value of the number of rows n is determined as the smallest integer exceeding m/k (step 420).

(4) Set the initial value "0" to the search result information line number counter i, check whether the search result information line number counter i is smaller than the number n of search result information lines, and check if i<n.
If so, perform the following processing (steps 430, 440)
.

(5) Set search result information line number counter i to “1”
"Counts up, sets the row number 252 of the next row in the memory buffer of the search result information row 250 to the row number of the search result information table 230 that stores the next search result information row 250, and also sets the row number sub The line number list 352 obtained in step 410 is divided into the maximum number of line numbers that can be stored in the line number sublist to create and set the i-th line number sublist, and the search result information line 250 is set in the list 260.
(Steps 440-460).

(6) Next, the line input/output processing unit 131 is called and the memory image of the search result information line created in step 460 is stored in the search result table 230 (step 470).

(7) The processes of steps 440 to 470 described above are repeatedly executed, and in step 440, if the value of the search result information line number counter i is equal to or greater than the number n of search result information lines, the search result information table is 230, the cursor name 321, the line number of the first line among the multiple search result information lines 250 stored in the search result information table 230, and the line stored in the search result table 230. The total number m of stored numbers is set, the current row number position 274 is initialized, and the process ends (step 4
80, 490).

Next, the processing operation of the row search control section 122 will be explained using the flowcharts shown in FIGS. 9 and 10.

(1) When an SQL command 340 that executes a row search from a cursor set by a cursor setting request 310 as shown in FIG. 6 is issued, cursor name 3
21, and the currently located actual data table 220
A search request 340 consisting of a retrieval offset 341 that specifies an offset value from the current line number position indicating the position of the line number in the line number list to the next line number to be positioned is passed to the line search control unit 122 ( Step 5
00).

(2) Search the cursor management information table 270 for an entry with the same cursor name 271 as the cursor name 321 passed in step 500, and find the first line number of the search result information line 250 from the entry. 2
72, the total number of stored row numbers 273 stored in the search result information table 230, and the current row number position 274 indicating the row number of the currently positioned actual data row 220 in the row number list. obtained (step 501).

(3) Add the passed offset 341 from the current line number position to the current line number position m obtained in step 501 to find a new current line number position m for which the search is to be performed (step 502).

(4) EOF (end of file) for the current line number position m obtained in step 502
If the check is performed and the current line number position m is EOF, the current line number position 274 in the cursor management information table 270 is set to (k+1) after the last line number position, and the line search control process is abnormally terminated. (Step 50
3-505).

(5) Perform an EOF check on the current line number position m obtained in step 502, and if the current line number position m is not EOF, the current line number position 274 in the cursor management information table 270 is The current line number position m obtained in step 502 is set (steps 503 and 506).

(6) From the current line number position m obtained in step 502, find the sequence number n of the search result information line 250 in which the current line number to be searched is stored, and set it in the search result line read number counter i. Initial value “1”
and set the row number r of the search result information table to be read to the initial value "j", the first row number of the search result information row 250.
is set (steps 507, 508).

(7) Call the line input/output processing unit 131,
From the search result information table 230, the search result information row 250 with row number r of the search result information table to be read obtained in step 508 and step 511, which will be described later, is read (step 5
09).

(8) Check whether the reading number counter i is smaller than the sequence number n of the search result information line of the current line number to be searched obtained in step 507, and if it is smaller, the read number counter i is checked in step 509. The value of the next row number 252 in the search result information row 250 is set to the row number r of the search result information table 230 to be read, the read number counter i is incremented by 1, and step 509
Return to step 510-512.

(9) In step 510, if the reading number counter i is greater than or equal to the sequence number n of the search result information line of the current line number to be searched obtained in step 507, the read number counter i in step 509 is In the search result information line 250, find the line number value p to be searched at the remainder position when the current line number position obtained in step 502 is divided by the maximum number of line numbers that can be stored in the line number sublist 260, The line input/output processing unit 131 is called to read the actual data row 220 with the row number p obtained above from the actual data table 200, and the process ends (steps 514 to 516).

[0044]

As explained above, according to the present invention, search result information is stored not in a dedicated file but in a DB table like the actual data, and the search result information is stored and reference input/output processing is performed. can be executed using the same means as for real data, making it possible to eliminate the need for a dedicated file management system, thereby reducing the scale of the database.

[Brief explanation of the drawing]

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

FIG. 2 is a diagram illustrating the data structure of an actual data table 200.

FIG. 3 is a diagram illustrating the data structure of a search result information table 230.

FIG. 4 is a diagram illustrating the data structure of a cursor management information table 270 managed in memory.

FIG. 5 is a diagram showing an actual data table that is an example of a search processing target.

FIG. 6 is a diagram illustrating a cursor setting request.

FIG. 7 is a diagram showing the relationship between a line number list obtained using a conditional term and a line number list obtained by combining using a logical operator.

FIG. 8 is a flowchart showing the processing of the cursor setting control unit 121.

FIG. 9 is a flowchart showing the processing of the row search control unit 122.

FIG. 10 is a flowchart showing the processing of the row search control unit 122.

FIG. 11 is a block diagram showing the system configuration of a relational database management system according to the prior art.

[Explanation of symbols]

100 Application (AP) process group 110
Relational database system 111 Transaction control unit 120 Data operation unit 130 Input/output control unit

Claims (3)

[Claims] 1. Means for generating and storing search result information by analyzing search conditions and executing a search; The relational database management system is characterized in that the search result information has the same structure as the actual data row and is stored in a database table having the same structure as the actual data table. Database management method. 2. The search result information is comprised of a sequence of information identifying rows such as line numbers of the actual data table, and the means for generating and storing the search result information is a cursor setting means. 2. The database management method according to claim 1. 3. The database management system according to claim 1, wherein the search result information stored in the database table is read using means for reading an actual data table.