JPH03246765A - Data retrieving system - Google Patents

Abstract

PURPOSE:To reduce a load to input a retrieval condition on a user by calculating and sorting similarity between unit information and that in a data base, and guiding the retrieval of the unit information designated by the user and that of similar unit information. CONSTITUTION:An editing means 10 edits the unit information by the designation of the user, and a similarity calculation means 6 calculates the similarity between the unit information designated by the user and that in the data base based on the degree of coincidence between a node and an arc, and a similarity storage table 8 stores a unit information name and the similarity. The similarity storage table 8 sorts the similarity based on the unit information name, and guides the retrieval of the unit information based on sorted similarity. Thereby, the load to input the retrieval condition on the user can be reduced.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a data retrieval method, and in particular to a database that performs data retrieval from a data group expressed by nodes and arcs connecting the nodes. The present invention relates to a data retrieval method that enables suitable retrieval of similar data when retrieving similar data based on data retrieval from .

(Prior Art) Conventionally, a method based on matching of data values has been proposed as a data search method (Japanese Patent Laid-Open No. 58103039:
Database system, JP-A-58-1370687 Data processing method, C, J, Date author: An 1ntr
production to database system
ems, Addison-ewesley Publ.
ishing Company + 1986) aThis method is suitable for retrieving simple data identified by data values, and is used for data retrieval applied in the application fields of office processing (e.g., personnel information management, inventory information management). It is a method.

The conventional data retrieval method was realized by a processing device having a configuration as shown in FIG.

That is, this processing device includes an input/output device W1 that inputs a data search request and displays search results, a CPU 2 that determines whether data matches, and a data storage device 5 that stores data in accordance with instructions from the CPU 2. It consists of a secondary storage control device 4 that reads and writes data, and a memory 3 that stores data from the secondary storage device 5. The CPU 2 is equipped with a data coincidence determination means 20, and the memory 3 is equipped with a data storage table 21.

Next, the operation of this conventional method will be explained with reference to the flowchart shown in FIG. The user is an input/output device
is operated to specify the file name to be searched (step 511). Thereby, the CPU 2 reads the data stored in the designated file (secondary storage device 5) under the control of the secondary storage control device 4, and transfers it to the memory 3 (step 512). Next, the user specifies the field name and data value as search conditions (Step 7).
13). As a result, the CPU 2 refers to the data storage table 21 in the memory 3 and extracts one record to be searched from the file transferred to the memory 3 (step 514), and determines whether the record satisfies the search conditions. is determined based on the match of the data values (step 515).

When the search conditions are satisfied, the record is output to the input/output device 1 (step 516), and the end of the search process for all records is determined (step 517).
). The processes of steps S14 to S17 are repeated until the search process is completed.

[Problem to be solved by the invention]

As mentioned above, conventional data search methods require the user to specify the field names and data values that make up a file as search conditions. There is a problem that the user must know the type, which increases the burden on the user, and the field of application of data retrieval is limited to the field of office processing where data values have direct meaning.

This invention was made to solve the above-mentioned problems, and it reduces the burden on the user of entering search conditions and also allows the user to know the structure of the file and the type of data stored in the field. Application fields where data values and their relationships have meaning (
The purpose of this invention is to obtain a data retrieval method that can efficiently perform data retrieval in the field of engineering, for example.

[Means to solve the problem]

The data search method according to the present invention includes an editing means 10 for editing unit information according to a user's specification, and a degree of similarity between a node and an arc to determine the similarity between the unit information specified by the user and the unit information in the database. and a similarity storage table 8 that stores the unit information name and the similarity, and calculates the similarity based on the unit information name. The information is sorted from the storage table 8, and the search for unit information is guided based on the sorted similarity.

[Effect]

The editing means IO edits the unit information according to the user's specifications. The similarity calculating means 6 calculates the similarity between the unit information specified by the user and the unit information in the database based on the degree of coincidence between nodes and arcs. The similarity storage table 8 stores the unit information name and the similarity. In other words, the unit information expressed by the node and arc specified by the user is used as a search condition, and the similarity between this unit information and the unit information in the database is calculated and sorted to find the unit specified by the user. Searching for information and similar unit information is guided.

〔Example〕

FIG. 1 is a block diagram showing the configuration of a processing device using a data search method according to an embodiment of the present invention. In FIG. 1, the processing device inputs a data search request;
Also, an input/output device 1 for displaying search results etc., and an editing means 1
0, a CPU (Central Processing Unit) 2 comprising similarity calculation means 6, etc., a memory 3 for storing a search pattern table 9, a connection data storage table 7, a similarity storage table 8, etc., and a unit information storage table 11. It is composed of a secondary storage device 5 that stores therein, etc., and a secondary storage control device 4 that controls this secondary storage device 5.

The editing means 10 edits the unit information according to the user's specifications. Unit information refers to data expressed by nodes and arcs connecting nodes. The similarity calculating means 6 calculates the similarity between the unit information specified by the user and the unit information in the database based on the degree of coincidence between nodes and arcs. The secondary storage control device 4 reads and writes data to and from the secondary storage device 5 according to instructions from the CPU 2. The connection data storage table 7 stores unit information expressed by nodes and arcs. The similarity storage table 8 stores the unit information name and the similarity. The search pattern table 9 stores unit information input by the user as search conditions. The unit information storage table 11 constitutes a data base of unit information.

FIG. 2 is a diagram showing an example of the unit information storage table 11.
The unit information expressed by the nodes and arcs shown in FIG. 3 is shown. FIG. 4(a) is a diagram showing an example of the search pattern table 9, and shows unit information expressed by the nodes and arcs shown in FIG.

Next, the operation of this embodiment will be explained with reference to the flowchart shown in FIG. First, in step S1, a search pattern input by the user by operating the input/output device 1 is
The CPU 2 stores the search pattern table 9 in the memory 3. For example, when unit information expressed by nodes and arcs shown in FIG. 5 is input from the input/output device 1 to the CPU 2 as a search pattern, the search pattern table 9 as shown in FIG. 3. In step S2, the unit information name is extracted from the unit information storage table 11 in the secondary storage device 5 and stored in the stack S. When the unit information storage table 11 has contents as shown in FIG. , stack S has unit information names A and B.
, C are stored.

In step S3, it is determined whether the stack S is empty, and if it is not empty, the unit information name is extracted from the stack S in step S4. In this case, the unit information name A is taken out from the stack S. Next, in step S5, the node names of the start point and end point corresponding to the unit information name are taken out from the unit information storage table 11 and used as the connection data time storage table 7. For example, if the node names of the start point and end point whose unit information name is A are extracted from the unit information storage table 11 shown in FIG. 2, the connection data time storage table 7 shown in FIG. 7 is obtained.

In step S6, the similarity calculation means 6 calculates the search pattern
The similarity between table 9 and connection data time storage table 7 is calculated. The process of step S6 is performed, for example, in FIG.
The similarity between the search pattern table 9 shown in a) and the connection data storage table 7 shown in FIG. 7 is carried out as follows. First, a variable that stores similarity, e.g. SI
Clear M to zero. Next, A1 in FIG. 7 is assigned to the starting node X in FIG. 4(a), and A2 is assigned to the ending node Y. Through this process, the contents of the search pattern table 9 become as shown in FIG. 4(b). If the data corresponding to the end node Z of this search pattern table 9 is in the connection data storage table 7, the similarity is increased by 1.
do. In the case of the connection data storage table 7 shown in FIG. 7, node A4 corresponds to node Z. Therefore, the value of the variable SIM that stores similarity becomes rlJ. Similarly, A2 is assigned to node X, and A3 is assigned to node Y.

In the case of FIG. 7, since node A5 corresponds to Z, the SIM
The value of is "2". A4. to node X. When assigning A5□A6, there is no data corresponding to node Z, so
Finally, the similarity between the search pattern table 9 shown in FIG. 4 (al) and the connection data storage table 7 shown in FIG. 7 is "2".

In step S7, the unit information name and similarity are stored in the similarity storage table 8. Next, the process returns to step S3, and it is determined whether the stack S is empty.

In this case, since the unit information names B and C are stored in the stack S, the unit information name is set to B and step S4. Step S5. Step S6°: Perform the process of step S7. Similarly, following the process in step S7, the determination process in step S3 is performed again. In this case, since the unit information name C is stored in the stack S, the unit information name is set to C and step S4. Step S5. Step S
6. Step S7 is processed. In the process of step S6, the node C1 of the connection data storage table 7 is connected to the node X in the search pattern table 9, and the node Y
Since the node C2 corresponds to the node C2 and the node C3 corresponds to the node Z, the similarity is "1". After the process in step S7, the determination process in step S3 is performed. In this case, since the stack S is empty, the process of step S8 is performed. Through the process of step S8, the data in the similarity storage table 8 is sorted in descending order of similarity. Figure 8 is the second
4 shows a similarity storage table 8 created from the unit information storage table 11 shown in the figure and the search pattern table 9 shown in FIG. 4(a).

In this way, the editing means 10 that edits the unit information expressed by nodes and arcs allows the unit information to be input easily, and the burden on the user to input search conditions is reduced. Furthermore, the similarity calculation means 6 that calculates the similarity between the unit information input as a search condition and the unit information in the database stored in the secondary storage device 5 allows the user to check the file structure and fields. It enables data retrieval without knowing the type of stored data, and further enables data retrieval in application fields (for example, engineering fields) where data values and their relationships are meaningful.

In the above embodiment, the unit information is expressed by the unit information storage table 11 composed of the unit information name, the starting node name, and the ending node name, as shown in FIG. A data structure other than a table format is acceptable as long as it can express the name and unit information structure (connection relationship between nodes and arcs), and the similarity between the search pattern entered by the user and the unit information. The method is not limited to the method in the above embodiment, but any method that uniquely calculates the similarity from the information that is the basis for calculating the similarity (information about the degree of coincidence between nodes and arcs) can be used. good. Furthermore, the process of sorting the data in the similarity storage table in descending order of similarity in step S8 of FIG. 6 can be replaced by a process of sorting it in ascending order, or the process of step S8 can be omitted. It is.

〔Effect of the invention〕

As described above, according to the present invention, the editing means edits the unit information according to the user's specification, and the similarity between the unit information specified by the user and the unit information in the database is determined by the matching between nodes and arcs. It has a similarity calculation means that calculates based on the degree, and a similarity storage table that stores unit information names and similarities, sorts the similarities from the similarity storage table based on the unit information names, and performs this sorting. to guide the search for unit information based on the similarities identified.
Unit information similar to the unit information specified by the user can be efficiently searched. This reduces the burden on the user to input search conditions, and also allows the user to easily understand the file structure and data stored in fields. It is possible to search for data without knowing the type, and furthermore, it is possible to efficiently search for data in applied fields where data values and their relationships have meaning (for example, the field of engineering).

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a processing device using a data search method according to an embodiment of the present invention, FIG. 2 is an example diagram showing the contents of a unit information storage table in this embodiment, and FIG. An example diagram of the unit information group in this embodiment, FIG. An example of a search pattern input by a user in this embodiment, FIG. 6 is a flowchart showing data search processing in this embodiment, and FIG. 7 is a connection data time storage table to be used for similarity calculation in this embodiment. FIG. 8 is an example diagram showing the contents of the similarity storage table in this embodiment. FIG. 9 is a block diagram showing the configuration of a processing device using a conventional data search method. FIG. 10 is a flowchart showing this conventional data search process. 2...CPU, 5...Secondary storage device,
6...Similarity calculation means, 8...Similarity storage table, 10...Editing means.

Claims (1)

[Claims] In a database whose unit information is data expressed by nodes and arcs connecting nodes, an editing means for editing unit information according to a user's specification, and a unit information specified by the user and the unit information in the database. has a similarity calculation means for calculating the similarity of the nodes and the arcs based on the degree of coincidence between the nodes and the arcs, and a similarity storage table that stores the unit information name and the similarity. A data search method characterized in that the similarity storage table is sorted based on the information name, and the search for unit information is guided based on the sorted similarity.