JP2001265803A - System and method for speeding up database access - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem that high-level knowledge is needed for establishing the know-how of access from an application program of basic operation, etc., to a database and for a design technology for obtaining a proper response at the time of access to an RDBM with an SQL statement and the problem that even SQL statements having been advanced in standardization need to be modified to the constitution of an application corresponding to special SQL statements prepared by RDBMSs when the latest function is fully used. SOLUTION: When the data of an RDBMS 1 are duplicated to an RDBMS 2, conversion information and access decision flags by items are provided and it is decided which of the RDBMSs is accessed. The RDBMS decided as the access object is accessed by conversion to a corresponding access indication and then the database access can be made fast without making client's sensing nor any alteration of the database access.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a business application which frequently accesses a database without changing programming know-how and without changing a database access portion of a business application which is already running. The present invention relates to a database access speed-up system and a database access speed-up method for speeding up the execution speed of database access.

[0002]

2. Description of the Related Art The speed of accessing a database is improving year by year, and recently, a method of accessing a relational database management system (RDBMS) by using an SQL statement is becoming mainstream. However, the establishment of know-how to access a database from an application program such as a core business and the design technique for obtaining an appropriate response often involve a high degree of knowledge.

[0003] Therefore, even in a core application group designed with a great deal of time, the speed of the database access portion may be a bottleneck, and an appropriate response may not be obtained, and it is often necessary to redesign the system. .

In recent years, there have been many cases where a large-scale database is provided separately from a database of a core business application such as a data warehouse construction.

In such a case, an operation of extracting data from the core business system for the data warehouse and making a copy of the data in the database of the data warehouse is performed. The application performs search aggregation processing and the like.

[0006] As described above, the database system on the server side is becoming complicated and diversified. In order to create an application in accordance with the characteristics of these databases in a business system, and to perform a design for drawing out the performance, It requires a lot of time and cost. Moreover,
As a method for accessing the database, an SQL statement is often used. However, since the specifications of the SQL statement differ depending on various RDBMSs and the description for obtaining the same result is different, the application is not changed. It was difficult to deal with various RDBMSs.

Japanese Patent Application Laid-Open No. H11-25116 discloses an optimal SQL statement automatic generation device and a storage medium for absorbing such differences in SQL statements and differences in RDBMS versions.

FIG. 12 shows an optimum SQL sentence automatic generation apparatus disclosed in Japanese Patent Application Laid-Open No. H11-25116. When an SQL sentence corresponding to the RDBMS 73 is input, the server process 74 searches the data storage unit 77 and sends it to the client 72. Pass the result. At this time, the SQL statement optimization process 75 searches the optimization rule storage unit 78 to extract the optimization rule and convert it into an SQL statement corresponding to the RDBMS 73.

[0009]

The above-mentioned conventional methods for accelerating database access have the following problems.

The standardization of database access methods such as SQL statements has been advanced, and a single business system has been using a plurality of diversified and distributed database systems.

However, even in standardized SQL statements, when using the latest functions, characteristic SQL statements are often prepared for each RDBMS.
It is necessary to change the creation of the application for each DBMS.

Further, recent RDBMSs are equipped with high-speed functions and special functions for differentiation, but applications must be changed in order to use these functions.

An object of the present invention is to adapt to various RDBMSs without changing existing applications.
By analyzing data access instructions and converting them into data access instructions tailored to the database, it is possible to effectively use multiple distributed database systems and further speed up database access, making business systems even easier to use. Is what you do.

[0014] Further, application service providers (AS
It is considered that a business system distributed through P) or the like accesses a centralized database, and it is assumed that a plurality of database systems such as a data center are centrally managed and various services are provided. . The present invention is intended for use in such fields.

In the data center, data of a plurality of different companies is stored in a database. Even if the database products are the same at first, some two-and-a-half years later, the database for each company will be different. For example, the product version and the presence or absence of correction information are examples. Also, when higher-speed hardware appears, it is necessary to provide means for switching to that hardware. As described above, the present invention also includes an element such as a software hub in a portion where the database devices are gathered.

[0016]

SUMMARY OF THE INVENTION A database access speed-up system according to the present invention is an access instruction acquisition means for receiving an access instruction from a client to a first relational database management system (RDBMS), and analyzes the access instruction. Second RDB
Access instruction conversion means for converting an access instruction suitable for an attribute of a database managed by the MS, the first RD
There is provided an access execution means for accessing the BMS and the second RDBMS and transferring one of the results returned to the client.

Further, an access instruction acquiring means for receiving an access instruction from a client to the first relational database management system (RDBMS), and analyzing the access instruction to determine whether to access the first RDBMS or the second RDBMS. Access instruction determining means for determining, an access instruction converting means for converting an access instruction suitable for an attribute of a database managed by the second RDBMS when determining access to the second RDBMS, the first RDBMS or the An access executing means for accessing the second RDBMS and transferring the result to the client is provided.

Further, the access execution means accesses the first RDBMS and the second RDBMS, combines the results, and passes the results to the client.

According to the method for accelerating database access according to the present invention, a step of acquiring an access instruction from a client program, a step of analyzing the access instruction for each lexical word to form a small sentence, and the first RDBMS managing from the access instruction Extracting a search sentence to the first database to be executed, and extracting the extracted search sentence to a second RD
Converting to a search sentence for a second database managed by the BMS, searching for the first database and searching for the second database using the converted search sentence, one of which has been previously returned And delivering the search result to the client program.

In the database access speed-up system according to the present invention, a step of acquiring an access instruction from a client program, a step of analyzing the access instruction for each lexical word and converting it into a small sentence, and the first RDBMS managing from the access instruction Extracting a search sentence to the first database, determining whether the extracted search sentence accesses the first RDBMS or the second RDBMS, and determining that the second RDBMS is accessed. Converting into a search sentence to a second database managed by the second RDBMS;
It comprises a step of searching the first database and a search of the second database based on the converted search sentence, and a step of combining the search results and passing the result to the client program.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1 Embodiment 1 of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing Embodiment 1 of the present invention. As shown in FIG. 1, in the database access speed-up system of the present embodiment, a server 1 includes an access instruction acquisition unit 3, an access instruction conversion unit 4, and an access execution unit 5, and the client 2 includes a server access unit 11 It has.

Referring to FIG. 1, the flow of the processing will be described. An access instruction from the database access unit 13 inside the client program 12 is passed to the server access unit 11 and requested from the access instruction acquisition unit 3. The access instruction acquisition unit 3 that has received the access instruction passes the access instruction to the access instruction conversion unit 4. The access instruction converting means 4 converts the access instruction for the database 1 into an access instruction suitable for the attributes of the database 2 by referring to the conversion information 8, and accesses the two access instructions for the database 1 and the access instruction for the database 2. Transfer to execution means 5.

The access execution means 5 issues an access instruction to each of the data body 7 of the database 1 and the data body 10 of the database 2 and waits for an access result. Then, only the result returned earlier is returned to the server access unit 11 on the client 2 side.

FIG. 2 is a block diagram of the conversion information 8 belonging to the database 1. Table name of database 1,
Item name, item data type, conversion table name converted to item data length and attribute corresponding to database 2, conversion item name, data type of conversion item, data length of conversion item, and target item can be converted There is provided a conversion target flag indicating whether or not this is the case. When the data replication to the database 2 is not performed normally or the data replication has not been performed yet, the conversion target flag is turned off. , Access to the database 2 can be temporarily interrupted.

FIG. 3 is a block diagram showing a processing procedure of the access instruction converting means 4. FIG. 4 is a block diagram showing a syntax tree which is analyzed for each lexical character and converted into small sentences when converting the access instruction. is there. As an example, the access instruction is
It is based on an SQL statement.

The conversion execution procedure will be described more specifically with reference to FIGS. Here, it is assumed that the data access instruction 23 issued from the client program 12 is SELECT SUM (A) FROM T1 in the SQL statement. This is a sentence for calculating the total of all rows for item A included in the table T1. This is access instruction acquisition means 3
When input to the access instruction converting means 4 through the above, the syntax is analyzed as shown in FIG. 4 and the memory (not shown) of the server 1
Expand the parse tree above. Information about the portion corresponding to the table and the portion corresponding to the item in this syntax tree is obtained from the conversion information 8 shown in FIG. 3 to determine whether the item can be converted to the item in the database 2, and in which table the item after the conversion exists Check whether the data type and data length are correct. At this time, in the syntax tree of FIG. 4, the table name is the table T1 of the small sentence 36, and the item name is the small sentence 3
Since it is known that the item A is No. 3, the conversion information 8 in FIG. 3 is referred to for the item A belonging to the table T1. Thus, it is found that the data type of the item A belonging to the table T1 is INT and the length is 4. The corresponding conversion item is item X belonging to table X1 and has a data type of INT and a data length of 4. Therefore, item A of table T1 matches item B of table X1 with both data type and data length. You can see that Thereby, the access instruction 23 is converted into SELECT SUM (X) FROM X1. Since conversion was completed normally, access instruction 23 to the original database 1 SELECT SUM (A) FROM
Access instruction for T1 and database 2 24 SELE
Transfer both CT SUM (X) FROM X1 to the access execution means 5. If the items in the database 1 do not match the items in the database 2 when referring to the conversion information 8, the conversion to the database 1 is not returned because a normal result is not returned even if the database 2 is accessed.
Only the QL statement is passed to the access execution means 5. Further, as described above, when the conversion target flag is turned off with reference to the conversion information 8, only the SQL statement for the database 1 is passed to the access execution unit 5 because the database 2 is not accessed.

FIG. 5, FIG. 6, FIG. 7, and FIG. 8 are flow charts showing the processing flow of the access instruction conversion means 4. In step S1 in FIG. 5, a data access instruction is obtained from the access instruction obtaining means 3, and in step S2, a syntax tree shown in FIG. 4 is created. Thereafter, in step S3, the conversion information 8 is expanded on a memory (not shown), and the syntax tree items are compared with the conversion information 8 to check whether the items in the database 1 can be converted into the items in the database 2. In step S4, a data access instruction corresponding to the item of the database 2 is generated.

FIG. 6 is a flowchart showing in detail the process of comparing the syntax and the conversion information shown in step S3 of FIG. 5 and converting the target data. In step S11, SELE is performed using the small sentence 31 at the head of the syntax tree shown in FIG.
It is determined whether the sentence is a CT sentence. If it is a SELECT sentence, it is treated as a conversion target, and other than the SELECT sentence is not converted here. Next, in step S12, the FROM phrase 3
5 is obtained and the table name in the database 1 is obtained. In step S13, it is checked whether or not there is an item in the syntax tree. If there is an item, the item name is obtained in step S14, and comparison with the conversion information 8 is performed in step S15. When there is no longer any item in the syntax tree shown in FIG. 4 in step S13, the process proceeds to step S16, and a SELECT statement corresponding to the database 2 is constructed.

FIG. 7 is a flowchart showing the details of the conversion item confirmation process in step S15 of FIG. In step S21, the table name indicated by the small sentence 36 and the item name indicated by the small sentence 33 in FIG. Step S
In step 22, the data type and length, which are the attributes of the target item, are extracted with reference to the conversion information 8. At this time, by referring to the conversion information 8 shown in FIG.
T and the length is found to be 4. Then, in step S23, the conversion item of the conversion information 8 is changed to step S23.
It is confirmed whether or not it matches the item extracted in 21.
Similarly, it is checked in step S24 whether the data types match, and in step S25, it is checked whether the data lengths match. If there is no match in this process, a normal result is not returned even if the database 2 is accessed, so the processing is terminated and no SQL statement is created for the database 2. When the item name, data type, and length match, the correspondence of the item is stored. As an example, when the SELECT statement shown in FIG. 3 is instructed, the fact that item A of table T1 can be converted to item X of table X1 is stored.

FIG. 8 is a flowchart showing in detail the process of constructing a SELECT statement corresponding to the database 2 in step S16 of FIG. Step S31
In step S27, it is checked whether or not there is an item name stored in step S27.
At 2, the item name is received. In step S33, the item name is replaced, and the process returns to step S31. When all the item names have been replaced, the process proceeds to step S34, where the table names are replaced, and a new SELECT statement SELECT SUM (X) FROM X1 is created.

In the present embodiment, even if the database 1 is a slow database, the effect is high if the database 2 is a database several times faster than that, and the database 1 and the database 2 can exist on different servers. is there. Further, the client program 12 does not need to know which database performs the database access and from which database the result is returned. Embodiment 2 Embodiment 2 of the present invention will be described with reference to the drawings. FIG. 9 is a block diagram showing Embodiment 2 of the present invention. As shown in the figure, the database access speed-up system according to the present embodiment includes an access instruction acquiring unit 3, an access instruction converting unit 4, an access executing unit 5, an access instruction determining unit 14 in a server 1, and a client 2 Is provided with server access means 11.

Referring to FIG. 9, the flow of the processing will be described. An access instruction from the database access unit 13 inside the client program 12 is requested from the server access unit 11 to the access instruction acquisition unit 3. Access instruction acquisition means 3 receiving the access instruction
Transfers the access instruction to the access instruction determination means 14. The access instruction determining unit 14 determines whether to access the database 1 by referring to the conversion information 8, to access the database 2, or to access both the database 1 and the database 2. When it is determined that the access to the database 2 is to be performed, the access instruction is passed to the access instruction conversion means 4 in order to convert the access instruction into an access instruction corresponding to the database 2. The access instruction conversion means 4
The access instruction to the database 1 is converted into an access instruction suitable for the attribute of the database 2 by referring to the conversion information 8, and the converted access instruction is returned to the access instruction determination unit 14. After that, the access instruction determining means 14 transfers one or two access instructions to the access executing means 5.

The access executing means 5 issues an access instruction to the data body 7 of the database 1 and the data body 10 of the database 2 in response to the passed access instruction, and waits for an access result. Then, the result is returned to the server access unit 11 on the client 2 side. When the result returned from the database 1 and the result returned from the database 2 are combined, the processing is performed here.

FIG. 2 is a block diagram of the conversion information 8 belonging to the database 1. Table name of database 1,
Item name, item data type, conversion table name converted to item data length and attribute corresponding to database 2, conversion item name, data type of conversion item, data length of conversion item, and target item can be converted There is provided a conversion target flag indicating whether or not this is the case. The difference from the first embodiment is that the conversion target flag takes into account information on whether only the database 1 is accessed, only the database 2 is accessed, or both are accessed. When data replication to database 2 is not performed normally or when data replication has not been performed yet, only database 1 is accessed to temporarily suspend access to database 2. So that
When data replication to the database 2 has been completely performed, it is possible to access only the database 2. If it is not clear which of the access result from the database 1 and the access result from the database 2 is returned first, it is possible to access both. That is, in the present embodiment, the access target is set to the database 1 in the conversion target flag.
Three states are set: database 2 and both database 1 and database 2.

FIG. 10 is a block diagram showing a processing procedure of the access instruction converting means 4. FIG. 11 is a block diagram showing a syntax tree which is analyzed for each lexical character and converted into small sentences when converting the access instruction. is there. The conversion execution procedure will be described more specifically with reference to FIGS. Here, the data access instruction 43 issued from the client program 12 is represented by an SQL statement SELECT SUM (A) SUM (B) FROM
Let it be T1. This is the item A included in the table T1.
This is a sentence for calculating the total of the entire rows for the item and item B. When this is input to the access instruction conversion means 4 via the access instruction acquisition means 3, as shown in FIG.
It parses into two syntaxes, SUM (A) FROM T1 and SELECT SUM (B) FROM T1, and expands them on the memory (not shown) of the server 1. A small sentence 56 corresponding to a table in this syntax tree,
Information on the small sentence 66 and the small sentences 53 and 63 corresponding to the items is obtained from the conversion information 8 to determine whether the information can be converted into the items in the database 2, which table the converted items exist in, and the data. Check that the type and data length are correct. At this time, the item B indicated by the small sentence 63 belonging to the table T1 of the database 1 indicated by the small sentence 66 has the same data type and data length as the item B of the table X1 of the database 2 by referring to the conversion information 8. Therefore, the syntax tree of SELECT SUM (B) FROM T1 is indicated by the data access instruction 45.
FROM X1. Since the conversion has been completed normally, both the access instruction 44 for the original database 1 and the access instruction 45 for the database 2 are passed to the access execution means 5. That is, at this time, the access instruction 43 for one database 1 is divided into the access instruction 44 for the database 1 and the access instruction 45 for the database 2.

The divided access instruction is issued from the access execution means 5 to the databases 1 and 2, and the returned results are combined by the access execution means 5 and returned to the server access means 11 on the client 2 side. .
Even if the database 1 is a slow database, the effect is high if the database 2 is a database several times faster than that, and the database 1 and the database 2 can exist on different servers. Further, the client program 12 does not need to know which database performs the database access and from which database the result is returned.

[0037]

Since the present invention is configured as described above, it has the following effects.

The access execution means accesses the first RDBMS and the second RDBMS, and returns the previously returned result to the client, thereby speeding up the access. Also, on the client side, which RDB
There is no need to sense whether an access has been made at the MS, and no change in database access means is required.

Further, the access instruction from the client to the first relational database management system (RDBMS) is analyzed to determine whether to access the first RDBMS or the second RDBMS. When it is determined that the user accesses the second RDBMS, the access instruction can be converted into an access instruction suitable for the attribute of the database managed by the second RDBMS, and the second RDBMS can be accessed. In this way, by distributing the access to different types of RDBMS, the access can be speeded up.

Further, the access execution means is a first RDBM
S and access to the second RDBMS, and the results can be combined and passed to the client.

[Brief description of the drawings]

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

FIG. 2 is a block diagram showing a configuration of conversion information 8 used in the first and second embodiments of the present invention.

FIG. 3 is a block diagram showing a processing flow in an access instruction conversion unit 4 according to the first embodiment of the present invention.

FIG. 4 is a block diagram showing a syntax tree obtained by analyzing an access instruction when processing is performed by access instruction conversion means 4 in the first embodiment of the present invention.

FIG. 5 is a flowchart showing a flow of an entire process in the access instruction converting means 4 according to the first embodiment of the present invention.

FIG. 6 is a flowchart showing a flow of comparison and conversion processing between a syntax and conversion information in the access instruction conversion unit 4 according to the first embodiment of the present invention.

FIG. 7 is a flowchart showing a flow of a process of confirming a conversion item in the access instruction converting means 4 according to the first embodiment of the present invention.

FIG. 8 is a flowchart showing a flow of processing for constructing a SELECT statement for the database 2 in the access instruction conversion means 4 according to the first embodiment of the present invention.

FIG. 9 is a block diagram showing an overall configuration of a second embodiment of the present invention.

FIG. 10 is a block diagram showing a processing flow in an access instruction conversion unit 4 according to the second embodiment of the present invention.

FIG. 11 is a block diagram showing a syntax tree obtained by analyzing an access instruction when processing is performed by an access instruction conversion unit 4 according to the second embodiment of the present invention.

FIG. 12 is a block diagram showing the overall configuration of a conventional database access system.

[Explanation of symbols]

 2 Client 3 Access instruction acquisition means 4 Access instruction conversion means 5 Access execution means 14 Access instruction determination means

Claims (5)

[Claims] 1. An access instruction acquisition unit for receiving an access instruction from a client to a first relational database management system (RDBMS), analyzing the access instruction and conforming to an attribute of a database managed by a second RDBMS. Access instruction conversion means for converting the first RDBMS and the second
A database access speed-up system comprising access execution means for accessing the RDBMS and transferring one of the results returned to the client to the client. 2. An access instruction acquisition means for receiving an access instruction from a client to a first relational database management system (RDBMS), analyzing the access instruction and determining whether to access the first RDBMS or the second RDBMS. Access instruction determining means for determining, an access instruction converting means for converting an access instruction suitable for an attribute of a database managed by the second RDBMS when determining access to the second RDBMS, the first RDBMS or the A database access speed-up system comprising access execution means for accessing a second RDBMS and transferring a result to the client. 3. The access execution means according to claim 1, wherein
3. The database access speed-up system according to claim 2, wherein an access is made to a BMS and said second RDBMS, and the results are combined and passed to said client. 4. A step of acquiring an access instruction from a client program, a step of analyzing the access instruction for each lexical sentence, and a step of searching from the access instruction to a first database managed by a first RDBMS. Extracting the extracted search sentence into a search sentence for the second database managed by the second RDBMS, searching for the first database, and converting the search sentence to the second database based on the converted search sentence. A method for speeding up database access, comprising a step of performing a search and a step of transferring one of the search results returned earlier to the client program. 5. A step of obtaining an access instruction from a client program, a step of analyzing the access instruction for each lexical sentence, and a step of retrieving a search sentence from the access instruction to a first database managed by a first RDBMS. Extracting the extracted search text, accessing the first RDBMS or the second RDBMS, and determining that the second RDBMS is accessed when the extracted search sentence is accessed by the second RDBMS. Converting to a search sentence to the database of 2,
A method of searching a first database and a search of a second database based on the converted search sentence, and a step of combining search results and passing the result to the client program. .