Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
2000-09-11
2004-06-29
Pardo, Thuy N. (Department: 2175)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C707S793000, C707S793000
Reexamination Certificate
active
06757670
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a computer system and more particularly to a method and a system for query processing, which process queries from users using one or more databases.
In recent years, there exist many database management systems (DBMS) in company's computer systems. From a historical point of view, those DBMSs have been added on a purpose base keeping pace with the development of the companies. Deregulation among different categories of industries is now rapidly progressing, which strengthens the tendency of the companies to add new operations to the existing ones. In this case, in response to the introduction of the new operations, an additional new DBMS is often introduced into the existing system.
On the other hand, each company is trying to provide the customers with better services including a new service differentiated from others, and as a result is trying to acquire a larger number of excellent customers. In such cases, because the company is more often required to analyze the company's activities, the customers' trends, and the like the data of which are accumulated in the DBMS group, the number of the companies that construct a data warehouse and a data mart are significantly increasing.
To construct the data warehouse and the data mart, it is necessary to put together data accumulated in the above-mentioned large number of DBMSs into one logically integrated database. In addition, beside the construction of database that is used as a base of analysis processing such as the data warehouse and the data mart, for the purpose of rapid starting of new operations, the existing DBMS groups are logically integrated to enhance competitiveness. It is because the logically integrated database as a base enables the company to construct application programs (applications) for the new operations in a shorter period of time.
Not only two or more DBMS groups within one company but also DBMS groups across two or more companies are now increasingly integrated. For example, a company group forming a group of companies integrates their information infrastructures to increase their profits as a whole; and companies in alliance with each other integrate their information infrastructures when they form a virtual alliance of the companies. Integration of the information infrastructures over a wider range provides applications developed on this integrated information infrastructure with commonality among a large number of divisions. As a result of it, overlap of application development can be avoided.
As a method for integrating the information infrastructures, there is a method of placing a “database hub” system, which provides access to the DBMS group in an integrated manner, between the DBMS group and an application group. In this method, the database hub receives a query (typically, a query described in Structured Query Language (SQL)) from an application, and then disassembles the query and converts it into an appropriate query to the DBMS group. After that, the database hub issues this disassembled and converted query to the DBMS group, collects data to create a result of the query from the DBMS, obtains the final result of the query received from the application, and returns the result to the application.
As a result of the integration of the information infrastructures, a variety of application queries are issued to the database hub. Those queries range over various kinds of queries such as a complex query for analysis operations, and a simple query for routine tasks. In particular, in the case of the complex query for analysis operations, the database hub and the optimization of the queries to the DBMS group concealed behind the database hub play an important role in performing the queries at high speed.
The optimization of the queries is broadly categorized into two groups: a rule-based query optimization and a cost-based query optimization. The cost-based query optimization, for example, is described in “Access Path Selection in a Relational Database Management System” written by P. G. Selinger and others (ACM SIGMOD Conference, May 1979, pp. 23-34; hereinafter referred to as Reference 1). In addition, those two kinds of optimizations are applied in a system (hereinafter referred to as Prior Art 1) described in “Oracle 8 Server Release 8.0, Overview: vol. 2” written by Oracle Corporation (pp. 19-7-9-11, Oracle Corporation, June 1997; hereinafter referred to as Reference 2). The cost-based query optimization provides the best query execution procedure (query execution plan) among various execution possibilities from the standpoint of execution cost.
Therefore, it is thought that the cost-based query optimization is suitable for the complex query. In fact, the cost-based query optimization is applied to many DBMS products, and enormous efforts are being expended to enhance it.
In reality, however, even if the cost-based query optimization is applied, the query is not always well optimized. As a result, the query is often processed by the execution plan that cannot be said to be the best for the user. There are two reasons for it.
Firstly, a query issued by an application is becoming complex at higher pace than that of technical enhancement of the cost-based optimization. Secondly, the cost-based optimization requires statistical information about a target database. However, there is a limit of details of the statistical information. Taking a relational database (a database which stores data in one or more tables consisting of one or more records having one or more columns) as an example, the following data are usually used as the statistical information: size of the table (number of records in the table); the minimum value, the maximum value, and an average value of values in the columns; number of different values in the columns; existence
onexistence of NULL value in the columns; and distribution of values in the columns. Regarding the cost-based optimization, the cost of a series of processing to obtain a result of the query is calculated using that statistical information. However, because the statistical information is merely data characteristics summary that is viewed from several aspects, it is difficult to estimate the correct cost of the series of processing completely.
To address this problem, in a system (hereinafter referred to as Prior Art 2) described in “Oracle 8 Server Release 8.0 Tuning” (pp. 8-8 to 8-27, Oracle Corporation, September 1997; hereinafter referred to as Reference 3), the application is allowed to direct a query optimization method to the query either by query unit or by user. However, because the direction can be given only by query unit or by user, it remains impossible to give an optimization direction in block to a group of queries that have been issued from different applications, and it also remains impossible to give the optimization direction in block to a group of queries that have common characteristics.
The following is an example: In the integration of the information infrastructures across two or more companies as described above, a database hub and a DBMS group under the database hub process queries which are issued from various organizations, divisions and posts, and which are related to operations with a variety of complexity and urgent necessity. For example, if two companies use a common application, and if there is a situation in which different optimization methods are required for optimization of a query from one company and for optimization of a query from another company, it remains impossible to address this case using the direction of optimization method by query unit or by user. Such differentiation of the optimization methods should not be realized by modifying the application. It is recommended that an administrator handling the database hub and the DBMS group under the database hub can control the differentiation in a batch.
In short, a first problem that is addressed by the present invention is to enable the administrator to control operation of the database hub a
Inohara Shigekazu
Nishizawa Itaru
Sagawa Nobutoshi
Shimizu Akira
LandOfFree
Method and system for query processing does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for query processing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for query processing will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3361207