JPH02109149A - Automatic knowledge extraction type data base control system - Google Patents

Abstract

PURPOSE:To prevent the destruction of the matching properties of data stored in a data base at addition, correction and deletion of the data base by producing the function dependency showing the characteristics of the data base by reference to the real data stored in the data base and storing the obtained function dependency. CONSTITUTION:An actual data base is taken out by a data base control system 102 based on a control program 108 of an automatic knowledge extraction type data base control system 109. Based on the data base, a knowledge production mechanism 103 produces automatically the function dependency showing the characteristics of the data base and stores it in a knowledge memory area 105 via a knowledge memory mechanism 104. Thus it is possible to prevent the destruction of the matching properties of data and at the same time to perform the addition, correction and deletion of the data base by reference to the storage contents of the area 105.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention extracts characteristics of data in a database using the data itself in the database, stores the extracted characteristics as knowledge about the database, It concerns how this knowledge can be used to maintain the integrity of data in a database.

(:3) [Conventional technology] Systems for maintaining the integrity of data in a database have conventionally existed, but in such conventional systems, users have no control over the nature of the data! - It was necessary to define constraints regarding the data in advance. (for example,
Ohm et al., Fundamentals of Database Systems, 1979).

[Problem to be solved by the invention]

Conventional database management systems have the burden of requiring the user to define constraints in advance to maintain data consistency, resulting in the problem of omissions or omissions in definitions. was there.

The purpose of the present invention is to automatically extract constraints from the data itself in a database, store the obtained constraints as knowledge about the data in the database, and use this knowledge to add or modify data to the database. - The purpose is to be able to detect erroneous data and abnormal data during deletion processing.

[Means to solve the problem]

In order to achieve the above purpose, we examine the data in the database and identify the functional dependencies that exist between data items (constraints regarding the relationship between data, such as the fact that if the circuit name is the same, the voltage is always the same) By finding this functional dependency and storing it as knowledge representing the properties (constraints) of the database, when data is added, modified, or deleted from the database, this knowledge can be used to This is to prevent the mixing of erroneous data and abnormal data.

[Effect]

In the present invention, since the knowledge represented by the properties of the database, that is, the functional dependencies, is created using the actual data itself in the database, it is possible to create error-free knowledge. The created knowledge is stored in the knowledge storage area, and when data is added, modified, or deleted from the database, the stored knowledge is referred to check whether the generated data is valid or not, and then the data is The data can be added, modified, or deleted without destroying the integrity of the data in the database.

〔Example〕

[Embodiment 1] FIG. 1 shows the overall configuration of the present invention. 1.01 is a database, 102 is a database management system, 10
3 is a knowledge creation mechanism for calling the data pace management system to read data in the database and searching for functional dependencies; 1.04 is a knowledge storage mechanism for storing the created knowledge; 105 is a knowledge storage mechanism for storing the created knowledge; A knowledge storage area 106 for storing data reads out necessary knowledge from the detection storage area 105 when data is added, modified, or deleted from the database, and detects data errors and errors.
The mechanism 107 for checking whether or not there is an abnormality is a mechanism for reading out the corresponding knowledge from the knowledge storage area 105 in accordance with a request from the knowledge error/abnormal data detection 95 mechanism 106. 1-
08 is a control program that receives a request from a user and transfers control to the database management system 1-02 and the knowledge creation mechanism 103.

109 is 102, 103, 104, 106°1. -0
This is an automatic knowledge extraction type database management system consisting of 7,108 systems. Reference numeral 110 denotes a terminal for issuing commands for adding, modifying, and deleting data to the database.

FIG. 2 shows an example of a database. 200 indicates a database in which sales information is recorded. 1.201 indicates item 14.
It shows that there are 5 item names: customer, phone number, 5 products, 1 price, and number of items sold.

202 to 207 are data 3, and 208 to 212 are functional dependencies existing in the database 200. Functional dependency is a term used to describe the nature of data in a database, such as product → price, where the price of a particular product is always the same.

For example, if we look at data 202, 204, and 206 for item X in the sales information database 200, the prices are all the same, 10 and ¥. In addition, there is a restaurant with 7 dishes - Dinner 2.
Looking at 03, 205, 207.

Both prices are the same at ¥20. Namely.

In the case of any product, the price is the same as long as the product is the same, so in the sales information database 200, the product →
There is a functional dependency called price. When the database 200 is examined, there is also a functional dependency relationship 209 of price→product. If the price is the same, the product is also the same.In other words, once the price is determined, the product is determined.
Although it is difficult to accept this assumption, this functional dependency exists in the sales information database 200, as it is clearly the case when looking at each data in the sales information database 200.

The functional dependence of 210 is this sales information database 200
can be understood immediately.

The functional dependency of 211 is customer, product→sales volume, which means that if the customer and product are the same, the sales volume is the same. Looking at the data for 204°206, the number of sales for both Company B and Product X is 50.

Also, looking at the 205°207 data festival, the number of sales for all seven products of Company B was 40. In the data of 202, company A, product It is not wrong to apply the functional dependence of numbers.

Therefore, in the data storage state of the sales information database 200, it is assumed that there exists a functional dependency of customers, products, and sales numbers. That is, customers and prices are functionally dependent on 212 →
This is because the sales quantity has a functional dependency of product → price, so it is possible to replace the product with price.

2, the functional dependencies of 208 to 212 are determined by a person looking at the sales information database 200, but in the present invention, the knowledge creation mechanism shown in 103 examines the data of 202 to 207 and automatically determines them. create.

FIG. 3 is a flowchart for automatically searching for functional dependencies.

The processing flow in FIG. 3 is as follows.

First, one of the items in the database is focused on and marked as ■ (301). Next, find one set of item names (one item name) from among the items excluding 1, and set this as J (
302).

It is checked whether the functional dependence, ■→I holds for the above ■ and J (303). When the functional dependency J→■ holds true, the functional dependency is stored (304).

Functional dependency, if ■→■ does not hold, use one of the other item name pairs (305), and if another item name pair exists, use that item name as a new pair J (306). ), the process returns to the above-mentioned process (303).

If another set of item names does not exist in the above process (305), the number of sets of item names for ■ is increased by one (307), and a set of item names is selected from among the item names excluding ■. Find one and make it a new ■ (308). If a set of item names exists in the process (308), the process returns to the process (303) described above.

If a set of item names does not exist in the process (308), another item name is found for ■ (309), and if another item name exists for the king, that item name is newly set to engineering. (310) and returns to the process (302).

In the process (309), it is determined that another item name exists for ■.
0) If not, the process ends.

FIG. 4 shows functional dependencies created according to the procedure shown in FIG. 3 stored in the knowledge storage area 105. If data is added or modified to the database,
Check the data with reference to the knowledge in Figure 4. If the data violates the knowledge content, a message will be sent to the user name.

In the case of the user's circuit where the data is i] (sh), the knowledge shown in FIG. 4 is re-created using the procedure shown in FIG. 3.

[Embodiment 2] The data 202 to 207 in FIG. 2 are examined, and the area shown in FIG. 5 is stored in the knowledge storage area. This process is performed by the knowledge creation mechanism 103.

[Embodiment 3] When there is a large amount of data in a database as shown in FIG. 2, knowledge is created for a part of the database rather than the whole. This is done by instructing to process X% of the total when instructing knowledge creation from the terminal 110.

〔Effect of the invention〕

According to the present invention, the system automatically creates database knowledge for maintaining database consistency by referring to the actual data in the database, so the database creator can There is no need to use drops in the mouth.9゜

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of a system 11 according to an embodiment of the present invention, FIG. 2 is a block diagram showing an example of a database, and FIG. 3 is a block diagram showing an example of a database. Processing flowchart, Figure 4,
FIG. 5 is a diagram showing an example of a storage format for storing created knowledge in a knowledge storage area. 1-03...Knowledge creation mechanism, 1.04...Knowledge storage area, 1,05...Knowledge storage area, 106 When adding, modifying, or deleting data to the database, check for erroneous and abnormal data. Mechanism to perform, 1.07 ・ChiimAe
A mechanism for extracting necessary knowledge from storage devices.

Claims (1)

[Claims] 1. In a computer system that manages a database,
A mechanism that examines data in a database to find functional dependencies (constraints regarding relationships between data) that exist between data items, a mechanism that stores this functional dependency as knowledge representing constraints of the database, and Knowledge characterized by having an area for storing knowledge, a mechanism for reading out this stored knowledge, and a mechanism for referring to knowledge and discovering erroneous data in the process of adding, modifying, and deleting data to a database. Automatic extraction type database management system. 2. A mechanism for examining data in a database to find the range of item data for a specific item name, a mechanism for storing data in that range as knowledge, an area for storing the knowledge, and a mechanism for reading this stored knowledge. An automatic knowledge extraction type database management system having a mechanism for extracting data and a mechanism for discovering abnormal data by referring to knowledge in the process of adding and correcting data to the database. 3. A mechanism that examines the data in the database and determines whether or not functional dependence holds for a specified amount of data or more, and determines that there is pseudo-functional dependence, and a mechanism that uses this pseudo-functional dependence as knowledge. A knowledge automatic extraction type database management system having a storage mechanism, an area for storing the knowledge, a mechanism for reading out the stored knowledge, and a mechanism for referring to the knowledge and discovering erroneous data. 4. In the system according to claims 1 and 3, each time data is added or modified, functional dependencies or pseudo-functional dependencies are searched for and stored as knowledge from a state where there is no data at all in the database. This is an automatic knowledge extraction type database management system that uses this knowledge to discover erroneous data when adding, modifying, or deleting data. 5. In the system according to claim 2, the value range of item data is checked and stored as knowledge every time data is added or modified, starting from a state where there is no data at all in the database, and this knowledge is used. This is an automatic knowledge extraction database management system that detects abnormal data during data addition and modification processes. 6. An automatic knowledge extraction type database management system having the content set forth in claims 1, 2, and 3 for a specified amount of data in a database. 7. A system that has a function of detecting functional dependencies between items in a data set with multiple items.