JP2003216618A - Data analysis device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve efficient data analysis by avoiding troublesome manual operations which must be done to edit a data table again when performing data analysis as to correlation among retrieval data from different databases. <P>SOLUTION: A data analysis device, which connects to a database server connected to a network to analyze the retrieval data from the databases, includes a database access control means for controlling access to the server; a retrieval data display means for displaying the data table obtained from each server; a data table creating means for creating a new data table by evaluating the data values of data items between the displayed data tables according to designated conditions to connect together a plurality of data tables; and a data table file preserving means for preserving the created data table in a file. <P>COPYRIGHT: (C)2003,JPO

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data analysis apparatus suitable for use in a data analysis software tool for performing an analysis by linking various data in a database connected by a network. It is about. 2. Description of the Related Art Generally, a database is used by being connected to a network. This allows database users to easily update data,
This is to enable search. As a result, regardless of the location and number of the databases, data can be accessed when necessary, and the database can be easily used from anywhere as long as the database has the right to access the database. [0003] In a relational database most currently used, data in the database is managed in units having a relation between the data (hereinafter referred to as a table). The table further includes a plurality of records. A record is a unit indicating the minimum data association consisting of data of a plurality of data items. When a data search is performed on one or a plurality of data items, data of other data items related to the data can be obtained as a search result at the same time. Within one database, the relationships between data items are organized as described above, and data duplication and update are managed without contradiction. Therefore, when analyzing a data group of search results obtained from one database, Excel (manufactured by Microsoft), 1
Graph creation, statistical calculation, and the like can be easily performed using spreadsheet software such as 2-2-3 (manufactured by Lotus). In this case, save the search results in a file once, after starting these software individually, read and analyze those files, or create an interface program for each spreadsheet software and database,
It is necessary to deliver the search results directly to these softwares on the memory for analysis. [0004] Obtaining search result data from a mere database is almost the same as obtaining data from a single database. In other words, the data of the search result is stored in a file or the data is delivered to spreadsheet software. However, the relationship between data items is not managed in a unified manner between data tables of different databases, so when analyzing the correlation of data items between search data tables obtained individually from them, the relationship is not You need to join the data tables through the data items you think are there. [0005] In this case, the association of the data items means that the attributes match, that the data analyst recognizes that there is a correlation, that the correlation is estimated, and so on. Therefore, each data value does not always match in the related data item. In addition, problems such as different data arrangements and duplicated data values occur, making it difficult to combine data tables. Therefore, before performing analysis with the spreadsheet software, the file was read individually into the spreadsheet software for re-editing the data, such as confirming the duplication of the above data values and defining the relationship between data items. It must be done manually later by the user. [0006] When performing data analysis using a plurality of databases under the above-described network, when data items of search data of different databases are analyzed in association with each other, individual analysis is performed. It is necessary to read out the search data storage file and create a program individually to create a new data table by associating specific data items between the files, or manually edit it, etc. There is a problem that complicated work is required. Therefore, an object of the present invention is to facilitate data analysis by connecting a plurality of data tables of different databases (the same database may be used) with flexibility and without specially creating a program. And According to the present invention, there is provided a data analyzing apparatus for connecting to a database server connected to a network and analyzing search data from the database. Database access control means for controlling data access, search data display means for displaying data tables obtained from each server, and by evaluating data values of data items between the displayed data tables under specified conditions. It is characterized by comprising a data table creating means for creating a new data table by combining a plurality of data tables, and a data table file saving means for saving the created data table as a file. According to the invention configured as described above, when combining a plurality of data tables of different databases, the data items between the data tables are designated without embedding the key items in each data table in advance. In addition, by specifying conditions for combining with the specified data, the data tables can be combined, so that data extracted from a plurality of databases can be easily analyzed. An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a main configuration of a data analysis tool implementing a data analysis device according to the present invention. FIG. 2 is a block diagram showing an example of a data record merging operation according to the present invention. FIG.
FIG. 4 is a block diagram showing an example of a connection condition specifying operation according to the present invention. As shown in FIG. 1, the data analysis tool of the present embodiment includes a database access control unit, a data table storage area, a search data display, a data table creation unit, and a data table file storage. The lines connecting the components in the figure indicate the flow of data and control. The database access control section controls data retrieval to various databases. It is possible to incorporate an interface unit for each database. The interface unit delivers the search data via a memory or the search data via a file to the tool. The data table storage area is an area for storing search results obtained from each database. The search data is stored in the main storage area for each database from the access control unit. This area is secured on a memory or a disk. Data is transmitted to and received from each unit in response to requests from the search data display unit, data table creation unit, and data table file storage unit. The search data display section displays search data from each database in a table format. Display search result data in a table with data item names. Thereby, it is possible to confirm the search result from each database. If the search data is inappropriate, the search is performed again via the database access control unit. Also, by displaying the data item name of each database, it is possible to confirm the data item name of the same definition. [0014] The data table creation section specifies a data table to be associated with each other by using the database name and the data item name in the search data display section, thereby creating a data table obtained by combining search tables of a plurality of databases into one. create. To create a data table, you first need to specify the join conditions for the records. Examples of the joining condition include joining records when the data values of the selected data items are equal to each other, joining records when the difference between the data is within 10% of the data value, and the like. Records satisfying such a joining condition are joined to create a new data table record. Thus, the contents of the created data table can be checked simultaneously with each data table via the search data display unit. If necessary, the newly created data table linked as described above is sent to the data table storage area, and is stored in the same area as one of the new data tables. FIG. 2A shows an example of a data table display in the search data display section. In this manner, the search result is displayed as a table in database units. The search result data table of the database A1 is
It consists of N records, and the search result data table from the database B1 consists of four records. In this example, the data item xn of the database A1
And the data item y2 of the database B1 have the same definition, and one data table can be created through this data item. In this embodiment, an example is shown in which records of the data tables are joined when the data of the data item xn and the data of the data item y2 match. By designating this combination condition in the data table creation unit, a new data table as shown in (3) of FIG. 2 is created. If there is no record that can be combined at the time of this combination, it is displayed as a missing value. By utilizing the database of records combined in this way, it becomes easy to investigate the correlation between the data item x1 and the data item yn. Here, three data records are effective for the correlation check. In the present embodiment, data items between different databases are used, but the same applies to different data tables of the same database. The same applies to three or more data tables. FIG. 3 shows an example of the connection condition specifying operation according to the present invention. This table specifies the database search result table name, data item name, and join conditions. Several join conditions can be set. Only one data item name for the join condition evaluation can be specified for each target table, but a plurality of target tables can be specified. A hatched line in the condition specification column indicates that no condition is specified. The result of FIG. 2 (3) is an example obtained as a result of the first combination condition of FIG. Database A1,
The join condition for B2 is the data item name xn
And a record in which the data of the data item y2 matches. Further, at the time of combining, if the corresponding record does not exist, it is processed as a missing value. Further, when the same record occurs, it means that the processing of not adding the duplication to the table is sequentially performed. The second join condition in FIG. 3 is to join records in which the difference between the three data items of the designated data items of the databases A2, B3, and C5 is less than 5, and if there is no corresponding record, the record is Does not add to the join table. As described above, the Japanese invention makes it possible to easily combine search data in a database or search data between different databases, and efficiently analyze data items. it can.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing a main configuration of a data analysis tool according to the present invention. FIG. 2 is a block diagram showing an example of a data record merging operation according to the present invention. FIG. 3 is a block diagram showing an example of a joining condition designating operation according to the present invention.

Claims (1)

Claims: 1. A data analysis device that connects to a database server connected to a network and analyzes search data from the database, comprising: a database access control unit that controls access to the server; Search data display means for displaying a data table obtained from the server; and combining a plurality of data tables by evaluating data values of data items between the displayed data tables under specified conditions, to form a new data table. And a data table file storing means for storing the generated data table as a file.