JPS6347826A - Managing method for data base - Google Patents

Abstract

PURPOSE:To enable keyword retrieval as to a data based upon a relational model structure by retrieving a value corresponding to a specified keyword as to respective attributes in an attribute value table and using a sum set of information having the same value as a retrieval result. CONSTITUTION:A sum set of values KW(1), KW(2),... as keywords is constituted in a domain table corresponding to (one or plural) optional attributes, and a keyword specified at the time of the storage of pieces of information Inf(1), Inf(2),... is set as an optional attribute value in the attribute value table 1 by referring the domain table 2(1). When information is retrieved, on the other hand, the value corresponding to the keyword is retrieved as to respective attributes (1)-(3),... in the attribute table 1 and the sum set of information having the same value as to the respective attribute which is obtained from the retrieval result is regarded as the retrieval result corresponding to the keyword.

Description

DETAILED DESCRIPTION OF THE INVENTION The invention relates to a relational model structure, specifically an attribute value table and a union of its permissible values for each attribute of this attribute value table. The present invention relates to a database management method based on a relational model structure having a domain table and a domain table.

[Prior Art] Recent technological innovations have led to the optimization of storage media, and this trend has become particularly noticeable in image file systems using write-once optical discs. Therefore,
Various search systems have been considered for searching information filed on such large-capacity optical discs, etc., but systems whose search database is based on a relational model structure have some disadvantages in terms of ease of data acquisition. It has various advantages from

Specifically, the database based on this relational model structure is as shown in FIG. 9, for example.

This basically consists of attribute value table 1 and domain table 2 (1), 2 (2), 2 (3), etc.
・It has the following. Flexible, sexual 1st table 1 is,
A key attribute and each attribute that is functionally dependent only on this key attribute (1
), (2), (3), . . . are set, and the structure is such that a value (attribute value) can be specified for each locality. In addition, each attribute in the above attribute value table 1 (
1), (2), (3)... The union of allowable values for domain table 2 (1), 2 (
2), 2 (3),... Specifically, the allowable values [A, B,
C, D, E, F,... ] (union) is added to domain table 2 (1) with allowable values (a,
b, c, d, e, f, ・ ] (union) is stored in the domain table 2 (2) as an allowable value [α,
β, γ, δ, ε.

ζ,... ] (union) is domain table 2 (3
) are composed of...

When constructing a database with the relational model structure described above, one piece of information (file) to be managed is assigned to each tupble (one horizontal row) in attribute value table 1, and that file is assigned to the key attribute. number for each attribute (1), (
2), (3), . . . specify 8 values to identify the file. For this 8-value designation, each domain table is referred to and the values allowed for each of the above attributes are used.

For example, in the case shown in Figure 9, file 1 attribute (1) = A, attribute (2) = b, attribute (3) = α, ・
...File 2 Attribute (1) = B, Attribute (2) = d, Attribute (3) = β,
...File 3 Attribute (1) - A, Attribute (2) = a, Tropism (3) = α, ・
A database is constructed by specifying values for each file, such as...

As mentioned above, a feature of a database having a relational model structure is that only one value can be taken for one attribute.

In addition, the above attribute value table 1 and each domain table 2
(1), 2 (2), 2 (3), . . . are generally created in an auxiliary storage device such as a magnetic disk in an actual system.

On the other hand, when searching a database with the above-mentioned relational model structure, a ``tabular format arbitrary specification search'' is the conventional optimal method. When performing this "tabular format arbitrary specified search", first, an index table is created based on the information in the attribute value table 1. For example, in the case of the database shown in FIG. 9, as shown in FIG. 10, index table 3 (1) for attribute (1), index table 3 (2) for fi property (2),
3), index table 3(3) is created respectively. This index table is created by sorting based on 8 values. For example, index table 3 (1
), for the value II A TL, file 1. File 3. . . . File 2 for the value “B 11.” File 4 for the value “’C,” etc.

For example, when searching for a file with attribute (1') = A and attribute (2) = a, the value II is retrieved from index file 3 (1).
A The file number of 11 is index file 3.
The file number of the value "aIT" is searched from (2), and the final result such as file 3 is obtained by performing AND processing on each search result.

In "Tabular format arbitrary specified search", the final search result is obtained by performing various processing such as AND processing of values between each attribute as described above, and OR processing of values within the same attribute. There is. In other words, in this "table format arbitrary specified search", 1
This search method is based on a database with a relational model structure in which only one value can be taken for each attribute.

[Problems to be Solved by the Invention] By the way, there is another search method called "keyword search," but this "keyword search" originally did not hold values for each attribute in each file, but instead stored multiple arbitrary keywords as keywords. This is a search method that targets databases that are maintained. Therefore, this "keyword search" cannot originally be applied to a database based on the relational model structure as described above.

On the other hand, in information filing systems and the like, diversity in information management is desired due to the diversity of information handled. That is, it is possible to change the management method for a database with a specific structure in various ways depending on the information to be handled.

Therefore, an object of the present invention is to enable a ``keyword search'' on this database based on a database based on a relational model structure that is originally considered optimal for ``table format arbitrary specified search.''

The specific technical background for achieving this task is that recent advances in semiconductor manufacturing description have made it easier to obtain large-capacity, low-cost semiconductor memories. There is.

[Means for Solving the Problems] As shown in FIG. 1, the present invention includes an attribute value table 1,
Each attribute (1) of this attribute value table 1.

(2), (3), ... A domain table 2 (1) that should constitute the union of its allowable values corresponding to
2 (2), 2 (3), ... The database management method is based on a relational model structure having the following. In such a database management method, the technical means for solving the above problem is to add the value KW as a keyword to a domain table corresponding to an arbitrary attribute (one or more), for example 2(1). (1),
Construct the union of KW(2),... and obtain information 1nr(
1), Inf(2), . Each attribute (1) in attribute value table 1 has a value corresponding to the specified keyword.

(2), (3), ... were searched, and the union of information having the same value for each attribute obtained from the search results was taken as the search result for the keyword.

[Operations] Specifically, when a database like the one shown in Figure 1 is constructed using the above method, for example, when searching for information on KW(3), the value KW(3) corresponding to this specified keyword is searched. ) for each attribute (1) of attribute value table 1,
Search for (2), (3), ... and the sum of information 1nf(2), Inf(4), ... having the same value KW(3) for each attribute obtained from the search results. The set becomes the search result for the keyword KW(3).

[Embodiments of the invention] Embodiments of the present invention will be described below based on the drawings.

FIG. 2 is a block diagram showing an example of the basic configuration of a filing system that implements the database management method according to the present invention.

Reference numeral 10 denotes a system control processing device that controls the entire system, and this system control processing device 10 is composed of devices such as a processor and a memory, like a general system. The system control processing device 10 controls an image input device 11 such as an OCR, an optical disc noodle machine 12, a magnetic disk device 13, an image output device 14 such as a laser printer, and also controls a display device 16, a keyboard device 17, etc. Information is exchanged with the configured user interface device 15.

Here, the information to be filed is the optical disk device 12.
At the same time, a search database of the stored information is constructed in the magnetic disk device 13. Further, search processing of the database built in the magnetic disk device 13, user interface processing, etc. are realized by software in the system control processing device 10.

First, a case will be described in which a "tabular format arbitrary specification search" is implemented.

For example, as shown in FIG. 3, assume a database containing a company's employee directory.

In this case, “employee number” is the key attribute, and the employee name is the key attribute.
, ``Age II'', ``Age II, and II Gender'' are attributes that are dependent on the key residence, respectively.Then, by creating an attribute value table as shown in Figure 3, the relational model structure can be created. A database is constructed.

This attribute value table is always stored in the magnetic disk device 13, but for example, when the system is powered on, this attribute value table is referred to and an index table as shown in FIG. 4 is created. This index table will be created on the main memory in the system VJ control processor 10, which is possible in this day and age where large capacity, low cost semiconductor memories are easily available. . It is sufficient for this index table to have information regarding the attributes "location", "age", and "gender". In this case, the key attribute "employee number" and the attribute "name" are not created in consideration of efficiency and the like.

Here, for example, the process of searching for an employee number where ``Affiliation'' = ``Sales'' and ``Gender'' = ``Male'' will be described.

In addition to the above-mentioned index table, the system has a bitmap memory in which employee numbers and pit positions are associated with each other for attributes such as "affiliation,""age," and "gender." Then, for the above condition specification, “
Referring to the index table corresponding to ``Department'' and ``Gender,'' on the corresponding bitmap memory, ``1 Department'' is placed in the position corresponding to the employee number having the value ``°Sales'' (Figure 5 (a). ), for ``gender'', ``male''
(See Figure 5(b))
, expand the bit "1" respectively. After that, AND processing (AND) of both bitmap memories is performed on another bitmap memory in which the employee number and bit position are similarly made to correspond (see Figure 5 (C)) a. As a result, the above conditions are met. Enter the employee number 11 i 11. It is decided that...

Next, assume that a "keyword search" is to be performed on a database that basically has the relational model structure described above.

First, as shown in FIG. 6 [b], the domain table contains the keywords ``'War Robber'' and ``Workstation'' that are permissible in the filing system.
, 'PC n, .

In addition, the 6th El[aJ, puff aisle number'' is set as key B property, and provisional attributes (hereinafter referred to as board bottom property) board bottom property (1), (2).

An attribute value table in which (3) is set has been created. The basic structure of this attribute value table is exactly the same as that shown in FIG.

In this state, if a keyword is specified when saving a file, for example, if the keywords ``powerbua'' and ``'OA'' are specified for file 1, the corresponding values are retrieved in the specified order by referring to the domain table above. Set "Word Processor" to bottom property (1) and "'OA" to board bottom property (2). If the order of specification is reversed, "○A 11" is set to board bottom property (2). 1) “’word processor”
is set to plate bottom property (2). In other words, the actual operator who is performing the input operation using the user interface device has the f
The attribute value table of MJ6 is not visible.

In this way, a database of attribute value tables as shown in FIG. 6 [aJ] is constructed.

Furthermore, in this case as well, each plate bottom property (1), (2).

Regarding (3), an index table as shown in FIG. 7 is created. The method of creating it is the same as the example in the library.

Here, for example, Keyword = ``word processor'' The following describes the process for searching for files.

In this case as well, as in the case of "tabular format arbitrary specified search", a bitmap memory is provided in which file numbers and pit positions are associated with each other for each plate bottom property (1), (2>, (3)). Then, in response to the above keyword specification, the index table corresponding to each board property (1), (2), and (3) is referred to, and a word processor is written on each corresponding bitmap memory. A pit "1" is developed in each position corresponding to the file number having the value "1".The state is as shown in FIG. The bitmap memory corresponding to bottom property (2) is shown in the same figure (b), and the bitmap memory corresponding to board bottom +1 (3) is shown in the same figure (c). OR processing (OR) is similarly performed on other bitmap memories in which the file numbers and bit positions correspond (see Figure 8(d)).As a result, the file number matching the above keyword specification is 1''.
, “2”, “6”, . . .

In the above embodiment, the data structure targeted by the r-table format arbitrary specification search J and the "keyword search" is shared as a relational model structure, and furthermore, the index table as the search module and the bitmap memory for search are made common. is planned.

Therefore, in this filing system, as long as a module related to conversion from search conditions to a search procedure and a module related to user interface are provided for each search method, it is possible to arbitrarily select "express format specified search" or "keyword search J." In other words, it becomes possible to easily select a search method according to the information to be stored.

In addition, in each of the above search processes, “
Using a Zabata bitmap memory that corresponds to the "employee number" or "file number" and the pit position, intermediate results of the search process, final results after AND processing, OR processing, etc. are stored in the bitmap memory. Since the information is expanded, the speed of the search process can be increased.

Furthermore, it is important to note that tree-based data, for example, information about companies, etc.
For data that is classified by section → division → section →..., if the database is made according to the relational model structure,
In addition to being functionally dependent on the key attribute and each attribute, each attribute itself is also transitionally dependent.

A search method suitable for a database with this type of structure is a "cruise search," but in this case, the database structure is basically a relational model structure, so the database structure and search module described above are common. In addition, it is easy to implement a "cruise search" in the system.

However, the modules related to the conversion from search specifications to search procedures and the modules related to the user interface must be suitable for the [Cruise Search J] as in the above case. .

[Effect of the invention 1 As explained above, according to the present invention, originally “
Regarding databases with a relational model structure for which ``table format arbitrary specification search'' is optimal, ``keyword search J is possible, so users can select an easy-to-use search method according to the information to be filed. search]
and "keyword search".

Even when the degree of freedom in selecting a search method is increased, the scale of the entire system does not increase unnecessarily because the database has a common relational model structure.

[Brief explanation of drawings]

Fig. 1 is a diagram showing a database applied to the method according to the present invention, Fig. 2 is a block diagram showing an example of a filing system in which the method according to the present invention is realized, and Fig. 3 is a diagram showing a table format arbitrary specified search. A diagram showing an example of a database created,
4 is a diagram showing an index table related to the database in FIG. 3, FIG. 5 is a diagram showing the state of the bitmap memory in the search process, FIG. 7 shows the index table related to the database in FIG. 6, FIG. 8 shows the state of the dot map memory in the search process, FIG. The figure is a diagram showing an index table interposed in the database of FIG. 9. [Explanation of symbols] 1... Attribute value table 2 (1, 2 (2), 2 (3)... Domain table 3 (1), 3 (2>, 3 (3)... Index table 10...System control 2il processing device 11...Image input device 12...Optical disk device 13...Magnetic disk device 14...Image output device 15...User interface device Patent applicant Fuji Xerox Co., Ltd. Company agent
Patent Attorney Tomohiro Nakamura (2 others) Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9, 1 Figure 10

Claims (1)

[Claims] A database management method based on a relational model structure having an attribute value table and a domain table that configures the union of allowable values corresponding to each attribute of this attribute value table, the method comprising: Configure the union of values as keywords in advance in the domain table corresponding to
When saving information, the keyword specified is set as the value of any attribute in the attribute value table by referring to the above domain table, while when searching for information, the value corresponding to the specified keyword is set as the value of each attribute in the attribute value table. 1. A database management method characterized in that a search result for the keyword is set as a union of information having the same value for each attribute obtained from the search results.