JP2012038066A - Data processor and data processing method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the efficiency of work for extracting columns having a correspondence relation between two pieces of two-dimensional data.SOLUTION: A delimitation dividing part 11 selects a column pair to be an analysis target in migration source data, and a correlation rule calculating part 12 selects a column pair to be an analysis target in migration destination data, calculates a support degree and certainty in each row in the column pair of the migration source data, and calculates a support degree and certainty in each row for each column pair of the migration destination data. A correlation differential value calculating part 13 performs a difference calculation of a support degree and certainty between rows in the migration source data, and performs a difference calculation of a support degree and certainty between rows in each column pair in the migration destination data, a comparison calculating part 14 performs a difference calculation between a differential value in the migration source data and a differential value in the migration destination data, and a determining part 15 determines a column pair of the migration destination corresponding to the column pair of the migration source on the basis of a result of the difference calculation.

Description

  The present invention relates to a technique for analyzing a correspondence relationship between columns when data is integrated, for example.

In data integration associated with system integration, it is necessary to establish column correspondence between tables in different databases and reflect the data contents in both.
Here, data integration is to realize data migration by eliminating design information differences such as column names and data arrangement in the migration source database and the migration destination database.
At this time, a technique for discriminating the correspondence between similar tables and columns between systems is called a schema matching technique.
As a basic method of the schema matching technique, there is an analysis method using schema information (column name, type, etc.) and instance information (word, value appearance pattern, etc.).
Further, as an applied technique, it is possible to discriminate the correspondence between a plurality of column sets.
Here, the correspondence relationship between a plurality of column sets means a correspondence relationship between a certain column set and another column set.
As one of the correspondence relationships between the plurality of column groups, there is a case where the column contents (data) at the migration source of system integration are divided at a specific position and then dealt with at a plurality of migration destination columns.
As a specific example, an example in which a phone number is handled in one column at the migration source, a phone number is divided into three columns at the migration destination, and the name is handled in one column as a name at the migration source. An example of dividing into first name and last name at the migration destination and handling them in two columns is given.

In the technique of Patent Document 1, the correspondence between a plurality of column sets is determined by calculating an association rule (support level / confidence level).
The correlation rule is the following two values indicating the correlation between a certain target A and target B.
The certainty factor is a probability that the A selector selects B.
The degree of support is the probability that A and B appear simultaneously in the entire relationship.
In other words, the certainty factor is the ratio of the number of records in which both the target A and the target B are included to the number of records in which the target A is included.
The support level is the ratio of the number of records in which both the target A and the target B are included to the total number of records.
A method for determining correspondence between a plurality of column sets according to Patent Document 1 will be described.
When attention is paid to two columns in the same table, when a certain column is designated, correlation rules are calculated with another column existing in the same table, thereby determining the correspondence between the two.
In the technique of Patent Document 1, for example, for the purpose of market research, the support degree and the certainty factor of the column relating to the purchase of wine and the column relating to the purchase of cheese, which are separate columns, are calculated and the wine is purchased. Correlations such as high probability that a person purchases cheese are extracted.

JP 2000-353163 A

According to the technique of Patent Document 1, it is considered that a high value is obtained for both the support level and the certainty level in the relationship (for example, surname and first name) obtained by dividing the same data into two.
However, in the case of data integration, there are a migration source and a migration destination, and column sets that should normally be dealt with are arranged in separate tables.
At this time, even if it is assumed that the number of records is the same, in the conventional method, the transfer source and the transfer destination have no relation in the order of record arrangement and are independent, and thus the correlation cannot be found.
For example, even if the column corresponding to “Last Name” at the migration source and the column corresponding to “First Name” at the migration destination are combined and the association rule calculation is performed, the combined data is stored in separate tables. It is the contents of existing data combined, and it is not data that exists in the same record, so there is not necessarily a specific relationship, and it is highly possible that a high value cannot be calculated by the correlation rule calculation, and judgment is impossible It is.
Further, if Patent Document 1 is used, it is only possible to obtain a conclusion that there is a high possibility that the name “Ichiro” of the destination data will appear for the last name of the source data: “Sato”. It is not possible to extract the migration destination data column corresponding to this column.

  The main object of the present invention is to solve the above-mentioned problems, and it is a main object of the present invention to improve the efficiency of the work of extracting a column having a correspondence relationship between two data.

The data processing apparatus according to the present invention
Select a column pair to be analyzed as the first column pair to be analyzed for two-dimensional first data that contains multiple fields and each field is divided into one of multiple columns. A column pair selection process execution unit for executing the column pair selection process to be performed;
A first appearance tendency analysis processing execution unit for executing a first appearance tendency analysis process for analyzing an appearance tendency of a concatenated field value obtained by concatenating field values of each column of the first analysis target column pair in units of rows;
For two-dimensional second data in which a plurality of fields are included and each field is divided into one of a plurality of columns, the column pair to be analyzed is set as 1 as the second analysis target column pair. Second appearance trend analysis processing for selecting the pair or more and analyzing the appearance tendency of the connected field value obtained by connecting the field values of each column of the second analysis target column pair in units of rows for each second analysis target column pair A second appearance tendency analysis processing execution unit for executing
The analysis result for the first analysis target column pair and the analysis result for each second analysis target column pair are analyzed, and the concatenated field value in the first analysis target column pair is determined for each second analysis target column pair. It has an approximation degree calculation process execution part which performs the approximation degree calculation process which calculates the approximation degree with an appearance tendency, It is characterized by the above-mentioned.

  According to the present invention, the appearance tendency of the connection field value of the first analysis target column pair of the first data and the appearance tendency of the connection field value for each second analysis target column pair of the second data are analyzed. Then, each analysis result is analyzed, and for each second analysis target column pair, the degree of approximation with the appearance tendency of the connected field value in the first analysis target column pair is calculated. It is possible to improve the efficiency of the operation of extracting the column pair in the second data that has a correspondence relationship with the second data.

FIG. 3 is a diagram illustrating a configuration example of a table integration device according to the first embodiment. The figure which shows the relationship between the table integration apparatus which concerns on Embodiment 1, and a database. FIG. 3 is a flowchart showing a processing outline of the table integration device according to the first embodiment. FIG. 4 is a diagram showing an example of column division according to the first embodiment. FIG. 6 is a diagram for explaining an association rule according to the first embodiment. FIG. 3 is a flowchart showing association rule calculation processing according to the first embodiment. FIG. 4 is a flowchart showing correlation difference value calculation processing and comparison calculation processing according to the first embodiment. FIG. 4 is a diagram illustrating an example of a support degree calculation result according to the first embodiment. FIG. 4 is a diagram illustrating a configuration example of a table integration device according to a second embodiment. FIG. 10 shows an example of column selection according to the third embodiment. FIG. 10 is a diagram illustrating a configuration example of a table integration device according to a fourth embodiment. FIG. 10 shows an example of column selection according to the fourth embodiment. FIG. 10 is a diagram illustrating a configuration example of a table integration device according to a fifth embodiment. The figure which shows the example of the transfer source data and transfer destination data which concern on Embodiment 1. FIG. The figure which shows the hardware structural example of the table integration apparatus which concerns on Embodiment 1-5.

Embodiment 1 FIG.
FIG. 1 shows a configuration example of a table integration device according to the present embodiment.
As shown in FIG. 1, the table integration device 1 is connected to a migration source database 502 of the migration source system 501 and a migration destination database 602 of the migration destination system 601.
In this embodiment, the migration source database 502 and the migration destination database 602 are relational databases.
In the table integration device 1 according to the present embodiment, the specific column of the two-dimensional data in the migration source database 502 (hereinafter referred to as migration source data) is the two-dimensional data in the migration destination database 602 (hereinafter referred to as migration destination data). It is determined which of the plurality of columns is in a correspondence relationship.
More specifically, in the table integration device 1 according to the present embodiment, the content obtained by dividing the data held in the specific column of the migration source data corresponds to any column of the plurality of columns of the migration destination data. It is determined whether it is in.

Although details will be described later, in the present embodiment, for example, migration source data and migration destination data shown in FIG. 14 are targeted.
FIG. 14A shows the migration source data, and FIG. 14B shows the migration destination data.
Both the migration source data and the migration destination data are two-dimensional data in which a plurality of fields are included and each field is divided into one of a plurality of columns.
In the present embodiment, the column of the system administrator of the migration source data is the analysis target.
Further, in the migration destination data, the individual name is shown in each column of the user, the use applicant, and the licensee.
For this reason, it is necessary to specify whether the column corresponding to the column of the system administrator of the migration source is a user, a use applicant, or a licensee.
In the migration destination data, the user represents a person who actually uses a predetermined system that is allowed to be used only when the use application is made and the use application is permitted.
The use applicant represents a person who has applied for the use of the system for the user. The user himself / herself may be used.
The licensed user is a person who has licensed the usage application for the system.
In the source data, the system administrator's "last name" and "first name" are stored in one column, but in the destination data, the personal name is stored in different columns of "last name" and "first name". .
Therefore, for comparison with the migration destination data, the column of the system administrator of the migration source data needs to be divided into a column representing “last name” and a column representing “first name”.
Here, a concatenation of column and column field values in units of rows is called a concatenated field value.
For example, a concatenated field value (for example, “Sato” + “Ichiro”) obtained by concatenating the column values representing “last name” and the column representing “first name” in a row unit after being divided in the migration source data. .
Similarly, the column value representing the “last name” and the column value representing the “first name” in the migration destination data concatenated in units of rows (for example, “Yamamoto” + “Ichiro”) is also referred to as a concatenated field value.
In the present embodiment, the migration source data is an example of the first data, and the migration destination data is an example of the second data.

As shown in FIG. 1, in the table integration device 1, the database connection unit 20 connects to the migration source database 502 and connects to the migration destination database 602.
As illustrated in FIG. 2, the database connection unit 20 loads connection information from the connection information holding unit 21 in the storage area 16, connects to the migration source database 502 in the migration source system 501, and migrates to the migration destination system 601. Connect to database 602.
Then, the database definition information 101, the instance data 102, and the delimiter information 103 are acquired, the acquired database definition information 101 is output to the definition information acquisition unit 17, the instance data 102 is output to the data acquisition unit 18, and the delimiter information 103 is stored. The data is output to the delimiter information acquisition unit 19.

The definition information acquisition unit 17 acquires the database definition information 101 from the database connection unit 20 and stores the acquired database definition information 101 in the definition information holding unit 161 in the storage area 16.
The database definition information 101 indicates, for example, the number of columns, the attribute of each column, the data type of each column, etc. for each of the migration source data and the migration destination data.
The target system from which the definition information acquisition unit 17 acquires the database definition information 101 may be configured such that a single system includes a plurality of databases.

The data acquisition unit 18 acquires the instance data 102 from the database connection unit 20 and stores the acquired instance data 102 in the acquired data holding unit 162 in the storage area 16.
The instance data 102 is a field value stored in the migration source data table and a field value stored in the migration destination data table.
Note that when the database connection unit 20 is not connected to the migration source database 502 and the migration destination database 602 via a network, the instance data may be acquired offline via a recording medium.

The delimiter information acquisition unit 19 acquires the delimiter information 103 from the database connection unit 20 and stores the acquired delimiter information 103 in the delimiter character information holding unit 163 in the storage area 16.
In this embodiment, an example in which the data value of the column of the migration source data is divided into two will be described. Therefore, the delimiter information 103 includes a mark and a mark when the data value of the column of the migration source data is divided into two. The delimiter is shown.

The delimiter division unit 11 divides the contents of a specific column of the migration source data based on the delimiter information 103 acquired by the delimiter information acquisition unit 19 and stores the contents in the divided data holding unit 164 as two types of virtual column data.
The column to be divided is designated by the user through the user I / F 22, for example.
The column pair (two virtual columns) stored in the divided data holding unit 164 by the delimiter dividing unit 11 is a column pair to be analyzed in the migration source data, and is an example of a first analysis target column pair. The column pair is also called a column set.
The delimiter dividing unit 11 performs a process of selecting the first analysis target column pair, and is an example of a column pair selection process executing unit.

The correlation rule calculation unit 12 obtains data from the divided data holding unit 164, performs a correlation rule calculation, and holds the result in the correlation rule calculation result holding unit 165.
More specifically, the correlation rule calculation unit 12 calculates the instance of the analysis target column pair (first analysis target column pair) of the migration source data, that is, the field value of each column of the analysis target column pair in units of rows. The support level and the certainty level are calculated for each linked field value linked in.
The support level and the certainty level represent the appearance tendency for each connected field value.
The support level and the certainty level are collectively referred to as an association rule.
Further, the association rule calculation unit 12 selects one or more column pairs to be analyzed in the migration destination data.
In this embodiment, combinations of all types of columns in the columns included in the migration destination data are targeted for analysis.
The column pair to be analyzed in the migration destination data is an example of the second analysis target column pair.
Then, the association rule calculation unit 12 concatenates the instance of the analysis target column pair (second analysis target column pair) of the migration destination data, that is, the connection of the field values of each column of the analysis target column pair in units of rows. Calculate support and confidence for each field value.
The support level and the certainty level in the migration destination data are calculated for each column pair.
The association rule calculation unit 12 is an example of a first appearance tendency analysis processing execution unit and a second appearance tendency analysis processing execution unit.

The correlation difference value calculation unit 13 calculates the difference value between the support levels obtained by the correlation rule calculation of the correlation rule calculation unit 12 and the difference value between the certainty factors for each of the migration source data and the migration destination data. The correlation difference calculation result holding unit 166 holds the difference value between the support level and the certainty level in the original data, and the difference value between the support level and the certainty level in the transfer destination data.
More specifically, the correlation difference value calculation unit 13 calculates a support level difference between linked field values in the column pair of the migration source data. Note that the difference in support level in the column pair of the migration source data corresponds to the first support level primary difference value.
Further, the correlation difference value calculation unit 13 calculates a difference in certainty between linked field values in the column pair of the migration source data. Note that the difference in the certainty factor in the column pair of the migration source data corresponds to the first certainty factor primary difference value.
The correlation difference value calculation unit 13 performs the same calculation for the transfer destination data.
That is, the correlation difference value calculation unit 13 calculates a support level difference between linked field values in a column pair for each column pair of migration destination data. Note that the difference in support level in the column pair of the migration destination data corresponds to a second support level primary difference value.
In addition, the correlation difference value calculation unit 13 calculates a certainty difference between linked field values in the column pair for each column pair of the migration destination data. In addition, the difference in the certainty factor in the column pair of the migration destination data corresponds to the second certainty factor primary difference value.
The correlation difference value calculation unit 13 then calculates the difference value of the support level (first support level primary difference value) and the difference value of the confidence level (first confidence level 1) in the transfer source data obtained in this way. Next difference value), the difference value of the support level (second support degree primary difference value) in the transfer destination data and the difference value of the confidence level (second confidence level primary difference value) are stored in the correlation difference calculation result holding unit 166. To store.
The correlation difference value calculation unit 13 is an example of an approximation degree calculation processing execution unit together with a comparison calculation unit 14 described later.

The comparison calculation unit 14 compares the correlation difference calculation results of the migration source and the migration destination, and writes the result, which is a correspondence candidate between column pairs, to the comparison calculation result holding unit 167.
More specifically, the comparison calculation unit 14 determines, for each column pair of the transfer destination data, the difference value (first support degree primary difference value of the support level) of the transfer source data calculated from the combination of the same concatenated field values. ) And the difference value of the support level of the transfer destination data (second support level primary difference value). The difference value between the support level difference value of the migration source data and the support level difference value of the transfer destination data corresponds to the support level secondary difference value.
The comparison calculation unit 14 performs the same calculation for the certainty factor.
That is, for each column pair of the migration destination data, the comparison calculation unit 14 calculates the difference value (first confidence primary difference value) of the certainty factor of the migration source data calculated from the same combination field value combination and the migration destination. The difference with the difference value of data certainty factor (second certainty factor primary difference value) is calculated. Note that the difference value between the certainty factor difference value of the migration source data and the certainty factor difference value of the migration destination data corresponds to the certainty factor secondary difference value.
Thereafter, the comparison calculation unit 14 performs the summation and quotient calculation for the calculated difference value of the support level, the summation and quotient calculation for the calculated difference value of the certainty factor, and then the support difference value and the quotient calculation after the quotient calculation. The calculation for integrating the difference values of the certainty factors later is performed, and the integration result is stored in the comparison calculation result holding unit 167.
The integration result represents the degree of approximation with the appearance tendency of the connected field value in the column pair of the migration source data for each column pair of the migration destination data.
The comparison calculation unit 14 is an example of an approximation degree calculation processing execution unit together with the correlation difference value calculation unit 13 described above.

The determination unit 15 is based on the integration result (approximation) of each column pair of the migration destination data calculated by the comparison calculation unit 14, and candidates for the column pair of the migration destination data having a relationship corresponding to the column pair of the migration source data. Are extracted as corresponding candidate column pairs.
The determination unit 15 is an example of a correspondence candidate extraction process execution unit.

In addition, the division | segmentation division | segmentation part 11, the correlation rule calculation part 12, the correlation difference value calculation part 13, the comparison calculation part 14, the determination part 15, the definition information acquisition part 17, the data acquisition part 18, the division | segmentation information acquisition part 19, and the database connection part 20 Can be programs that implement the processes described above.
When each element is a program, a CPU (Central Processing Unit) (not shown) executes the program for each element, and the above-described processing is performed.

  The storage area 16 is a data storage area realized by a memory or a hard disk.

Next, the operation will be described.
FIG. 3 is a flowchart showing an outline of processing of the table integration device 1 shown in FIG.
Initially, the process outline | summary of the table integration apparatus 1 is demonstrated along the flowchart shown in FIG.

First, in step S <b> 1, the definition information acquisition unit 17 acquires the column name / column order of the target table via the database connection unit 20 and stores it in the definition information holding unit 161.
In addition, the data acquisition unit 18 acquires data of the migration source database 502 and the migration destination database 602 via the database connection unit 20 and stores the acquired data in the acquired data holding unit 162.
The delimiter information acquisition unit 19 receives a specification of a column to be analyzed from the user via the user I / F 22, acquires delimiter information via the database connection unit 20, and stores delimiter information Store in the unit 163.
Subsequently, the delimiter division unit 11 acquires the migration source data from the acquired data holding unit 162, and divides the character string (field value) in the analysis target column designated by the user at the acquired delimiter character information portion. Then, the front part and the rear part not including the divided part are separately stored in the divided data holding unit 164 (column pair selection process).
As described above, the two columns after being divided by the divider / divider 11 correspond to the first analysis target column pair.

  Next, in step S2, the correlation rule calculation unit 12 performs a correlation rule calculation on both the migration source data and the migration destination data (first appearance tendency analysis process) (second appearance tendency analysis process).

Next, in step S3, the correlation difference value calculation unit 13 performs correlation difference value calculation using the result of the correlation rule calculation (approximation calculation process).
Next, in step S4, the comparison calculation unit 14 uses the correlation difference value calculation result to perform comparison calculation between the transfer source and the transfer destination (approximation calculation process).

Finally, in step S5, the determination unit 15 determines the correspondence between the migration source column and the migration destination column set, and outputs the result.
As an output method, file output, module output, interface, and the like are assumed.

  Hereinafter, details of steps S1 to S4 will be described.

First, step S1 will be described.
The definition information acquisition unit 17 acquires the column name / column order of the target table via the database connection unit 20 and stores it in the definition information holding unit 161.
Further, the data acquisition unit 18 acquires the migration source data from the migration source database 502 via the database connection unit 20, acquires the migration destination data from the migration destination database 602, and stores it in the acquired data holding unit 162. .
Further, the delimiter information acquisition unit 19 receives the specification of the column to be analyzed from the user via the user I / F 22 and acquires delimiter information via the database connection unit 20.
Subsequently, the delimiter division unit 11 acquires the migration source data from the acquired data holding unit 162, and divides the character string (field value) in the analysis target column designated by the user at the acquired delimiter character information portion. Then, the front part and the rear part not including the divided part are stored in the divided data holding unit 164 separately.
As a specific example of delimiter division, consider a case where “” (double-byte space) is given as a delimiter as shown in FIG.
At this time, with respect to the “name” column in the migration source data, the first and last full-width spaces are separated, and the portions before and after the matching portions are held as separate virtual columns.

Next, in step S2, the correlation rule calculation unit 12 calculates the support level and the certainty level, which are correlation rules.
Here, the correlation rule indicates a correlation between a target A and a target B by the following two values.
Confidence: Probability that A selector chooses B Support: Percentage of the whole relationship (A and B appear at the same time)
As described above, the certainty factor is the ratio of the number of records in which both the target A and the target B are included to the number of records in which the target A is included.
The support level is the ratio of the number of records in which both the target A and the target B are included to the total number of records.

A calculation example of the association rule will be described with reference to FIG.
For the concatenated field value “Sato → Ichiro” in the figure, there are two sets of “Ichiro Sato” for the total number of columns 5, so the support level is calculated as 2/5 = 0.4.
In addition, with respect to the concatenated field value “Sato → Ichiro”, there are two “Ichiro” for the total number of “Sato”, and therefore the certainty factor is calculated as 2/4 = 0.5.
The association rule calculation unit 12 calculates the support level and the certainty factor only for the analysis target column for the migration source data.
For example, in the case of the migration source data in FIG. 14A, the support level and the certainty factor are calculated for each of the concatenations of the field values in the same row in the two columns after dividing the name of the system administrator.
In the example of FIG. 14A, the support level and the certainty factor are calculated for each of “Sato → Ichiro”, “Sato → Jiro”, “Sato → Saburo”, and “Suzuki → Ichiro”.
On the other hand, for the migration destination data, the association rule calculation unit 12 calculates the support level and the certainty factor for each connection of field values in the same row for all types of column combinations.
For example, in the case of the migration source data in FIG. 14B, the combination of the user's “last name” column and the “first name” column (“Yamamoto → Ichiro”, “Watanabe → Saburo”, etc.), Combinations of “last name” and “first name” columns (“Ota → Mitsu”, “Suzuki → Junko”, etc.), and combinations of “Last name” and “First name” columns of the licensee (“Sato → Ichiro”) ”,“ Suzuki → Shiro ”, etc.), the combination of the user ’s“ last name ”column and the user ’s“ first name ”column (“ Yamamoto → Mitsu ”,“ Watanabe → Junko ”, etc.), A combination of the “last name” column and the “name” column of the licensee (“Yamamoto → Ichiro”, “Watanabe → Ichiro”, etc.), the user's “last name” column and the user ’s “first name” column Combinations (“Ota → Ichiro”, “Suzuki → Saburo”, etc.), the combination of the “last name” column of the applicant for use and the “name” column of the licensee (“Ota → Ichi” ”,“ Suzuki → Ichiro ”, etc.), the combination of the“ Last name ”column of the licensee and the“ First name ”column of the user (“ Sato → Ichiro ”,“ Sato → Saburo ”, etc.), The support level and the certainty factor are also calculated for combinations of the “last name” column and the “first name” column of the application applicant (“Sato → Jun”, “Sato → Junko”, etc.).
For example, if the migration destination data in FIG. 14B includes a column such as use date and time, the combination of the user's last name and use date and time values (for example, “Yamamoto → July 10, 2010”, etc.) Calculate support and confidence.
In addition, since it is clear that such a combination does not correspond to the analysis target column (combination of "last name" and "first name") in the source data, the combination of "first name" and "first name" in the destination data You may perform the setting which makes only the object of calculation of support and reliability.

  As described above, the association rule calculation unit 12 calculates a support / confidence pair for the analysis target column pair for the migration source data, and supports / confidence for all data combinations of the column pair for the transfer destination data. The set of degrees is calculated, and the calculation result is stored in the correlation rule calculation result holding unit 165.

  Next, the correlation rule calculation in step S2 will be described in detail with reference to the flowchart of FIG.

First, in step S201, reading of column contents is started.
Subsequently, in step S202, an instance corresponding to the divided front portion is read.
This corresponds to the data of the “last name” column in the example of FIG. 5 (FIG. 14A).
Subsequently, in step S203, comparison with the storage area is started, and it is confirmed whether or not the data read at the present time exists in the storage area.
In step S205, if the read data exists, the internal management variable is counted up by one.
If the read data does not exist, the data is registered in the internal storage area in step S206.
Here, data registration refers to registering data in the internal storage area and associating the number serving as an index, thereby leading to improved search performance in the migration step.
Subsequently, in steps S207 to S211, the instance of the rear part is read, and the same processing as that of the front part is performed.
Subsequently, in step S212, it is determined whether all instances have been read. If not all have been read, the next record in FIG. 5 (FIG. 14A) is read.
Specifically, reading of the front part and reading of the rear part corresponding to steps S202 to S211 are performed.
When reading of the divided data into the divided data holding unit 164 is completed, support level calculation / sorting is performed on the read data in step S213 to step S217, and subsequently in the correlation table format (FIG. 8). Thus, the writing to the correlation rule calculation result holding unit 165 is performed.
For the correlation rule calculation for the migration destination data, the flow shown in FIG. 6 is performed for each combination of columns included in the migration destination data.

FIG. 8 shows an example of the support level calculated for the migration source data in FIG. 14A and the migration destination data in FIG.
For the migration destination data, the combination of the user's "last name" column and the "first name" column ((A) in FIG. 8), the user name of the user name "last name" and the "first name" for reasons of drawing. Column combination (FIG. 8B), only the combination of the “last name” column and the “first name” column of the licensee (FIG. 8C), Support for all column combinations is included.
The certainty factor is also managed in the same format as in FIG.

Next, details of the calculation of the correlation difference value calculation unit 13 in step S3 and the calculation of the comparison calculation unit 14 in step S4 will be described.
In step S3 and step S4, the calculation of the following equation (1) is performed on the correlation rule calculation result obtained in step 2 to calculate a correlation comparison intermediate result.
In the following formula (1), ai is the i-th numerical value in the correlation table of the migration source data (for example, the numerical value of the i-th support degree when the support levels of “last name → first name” are arranged in descending order). It is.
bi is a numerical value for the same character string as the character string (combination of last name and first name) corresponding to ai in the migration destination data.
For example, in the case of FIG. 8, when i = 1, ai is a value for “Sato → Ichiro” and is 0.2, and the migration destination data is the first “last name → first name”. “Sato → Ichiro” does not exist in the column pair, bi is 0, “Sato → Ichiro” does not exist in the second “last name → first name” column pair, and bi is 0. In the third “last name → first name” column pair, “Sato → Ichiro” exists, and bi is 0.2.

  Note that the correlation comparison intermediate result is calculated for both the support level and the certainty level.

The details of step S3 and step S4 will be described below based on the flowchart of FIG.
When the flow of FIG. 7 is executed, the calculation of the above equation (1) is performed.

The correlation difference value calculation unit 13 calculates a correlation value difference when capturing two column sets in units of records in both the migration source data and the migration destination data with respect to the support level and the certainty level.
In step S301, the correlation difference value calculation unit 13 prepares the entire correlation calculation result of the migration source data in a form that can be calculated by the correlation rule calculation result holding unit 165.
Subsequently, in step S302, the correlation difference value calculation unit 13 performs a difference calculation on the migration source data for the combination of correlation values.
In the difference calculation, a difference value with respect to another row is obtained for each row of the migration source data.
For example, assuming that the correlation table of the migration source data is composed of three rows “Sato → Ichiro”, “Sato → Jiro”, and “Sato → Saburo” in FIG. 8, the first row (“Sato → Ichiro”). ) And the second line ("Sato → Jiro") support value difference, the second line ("Sato → Jiro") support level and the third line ("Sato → Saburo") support level The difference value between the support level of the first line (“Sato → Ichiro”) and the support level of the third line (“Sato → Saburo”) is calculated.
Also, the difference between the certainty of the first line (“Sato → Jiro”) and the certainty of the second line (“Sato → Jiro”), the certainty of the second line (“Sato → Jiro”) and the third line The difference between the certainty of the eyes ("Sato → Saburo") and the certainty of the first line ("Sato → Ichiro") and the third line ("Sato → Saburo") are calculated The

Subsequently, in step S303, the correlation difference value calculation unit 13 selects one column set (for example, (A) in FIG. 8: “Yamamoto → Ichiro”) in the correlation table of the migration destination data. Each row of the column set starting with “0.2” is read from the correlation rule calculation result holding unit 165, and in step S304, the row used for calculating the correlation difference value from the correlation table of the migration source data is read. One of the combinations (for example, the combination of “Sato → Ichiro” and “Sato → Jiro” in FIG. 8) is read.
Next, in steps S305 and S306, the correlation difference value calculation unit 13 selects the combination of rows read in step S304 (for example, "" in FIG. 8) in each row of the column set of the correlation table of the migration destination data read in S303. (Sato → Ichiro) and “Sato → Jiro” combination) is searched, and if there is a combination of the same row, the correlation difference for the same combination as the source data on the destination data side Subtraction between values is performed (step S307).
At this time, if the row (for example, “Sato → Ichiro” in FIG. 8) used for the difference value calculation for the migration source data does not appear in the correlation table of the migration destination data read in S303, Calculates the difference value by assigning 0 to both support and certainty.

In step S308, the correlation difference value calculation unit 13 determines whether or not the last set (for example, the combination of “Sato → Ichiro” and “Sato → Saburo” in FIG. 8) has been read from the correlation table on the source data side. If not, the subsequent set is read in step S304. If the reading is completed, the comparison calculation unit 14 performs comparison calculation in step S309.
Details of the comparison calculation in S309 will be described later.
In step S310, it is determined whether the reading / calculation in S303 to S309 has been completed for all the patterns in the column set in the correlation table on the migration destination data side. If not, another candidate column set (for example, FIG. (B): Read each row of the column starting with “Ota → Miru” and the column starting with “0.2”).
When all the reading is completed, the process is completed.

Next, a specific example of calculation in step S3 for realizing correlation difference value calculation and step S4 for realizing comparison calculation will be described with reference to FIG.
Here, the description will be made on the assumption that the correlation table of FIG.
In the following, the support level will be described, but the same applies to the certainty level.

(1) The correlation difference value calculation unit 13 obtains the following pairs of distances from the migration source data and uses them as indices indicating the order relationship.
1st and 2nd difference (support level of “Sato → Ichiro” −support level of “Sato → Jiro”) = 0.1
Difference between second and third (support level of “Sato → Jiro” −support level of “Sato → Saburo”) = 0.05
1st and 3rd difference (support level of “Sato → Ichiro” −support level of “Sato → Saburo”) = 0.15

(2) Moreover, the correlation difference value calculation part 13 calculates | requires a distance similarly with respect to the group calculated | required by said (1) for every column pair of transfer destination data.
(A) Column (A) in the migration destination data of FIG. 8 (support level of “Sato → Ichiro” −support level of “Sato → Jiro”) = 0
(Support level of “Sato → Jiro” −Support level of “Sato → Saburo”) = 0
(Support level of “Sato → Ichiro” −Support level of “Sato → Saburo”) = 0
(B) Column (B) in the migration destination data in FIG. 8 (support level of “Sato → Ichiro” −support level of “Sato → Jiro”) = 0
(Support level of “Sato → Jiro” −Support level of “Sato → Saburo”) = 0
(Support level of “Sato → Ichiro” −Support level of “Sato → Saburo”) = 0
(C) Column (C) in the migration destination data of FIG. 8 (support degree of “Sato → Ichiro” −support degree of “Sato → Jiro”) = 0.15
(Support level of “Sato → Jiro” −Support level of “Sato → Saburo”) = 0.05
(Support level of “Sato → Ichiro” −Support level of “Sato → Saburo”) = 0.2
In (A) and (B) of the calculation on the migration destination side, since each column does not include “Sato → Ichiro”, “Sato → Jiro”, and “Sato → Saburo”, each support level is set to 0. Difference calculation is performed.
In addition, in (C) of the migration destination side calculation, “Sato → Saburo” does not exist in the column, so the difference calculation is performed assuming that the support degree of “Sato → Saburo” is 0.
That is, the value of the character string that exists in the migration source and does not exist in the migration destination is regarded as 0 and is calculated.

The comparison calculation unit 14 subtracts the difference values of (1) from the difference values of (2) (A) to (C) for the same row combination, sums the subtraction values of each row, and calculates the absolute value of the total value. Is taken and divided by the number of matching combinations to obtain a distance value related to correlation (intermediate correlation comparison result).
The specific calculation result of the above example is as follows.

[(2) (A) difference value]-[(1) difference value] and [(2) (B) difference value]-[(1) difference value]
(Support level of “Sato → Ichiro” −Support level of “Sato → Jiro”) = − 0.1
(Support level of “Sato → Jiro” −Support level of “Sato → Saburo”) = − 0.05
(Support level of “Sato → Ichiro” −Support level of “Sato → Saburo”) = − 0.15
Absolute value of total value = 0.3
Quotient calculation value = 0.1

[(2) (difference value of (C)]-[(difference value of (1)]]
(Support level of “Sato → Ichiro” −Support level of “Sato → Jiro”) = 0.05
(Support level of “Sato → Jiro” −Support level of “Sato → Saburo”) = 0
(Support level of “Sato → Ichiro” −Support level of “Sato → Saburo”) = 0.05
Absolute value of total value = 0.1
Quotient calculation value ≒ 0.03

As described above, the above calculation is performed for both the support level and the certainty level.
Further, the correlation comparison intermediate results of the support level and the certainty level are integrated by the following formula using the variable α that can be specified by the comparison calculation unit 14.
Integrated result = α × (intermediate comparison result of support level) + (1−α) × (intermediate comparison result of confidence level)
However, the formula for obtaining the integration result may be in another form.
The integration result is stored in the comparison calculation result holding unit 167.

Finally, in step S5, the determination unit 15 determines a migration destination column set corresponding to the migration source column based on the integration result, and outputs the determination result.
The determination method includes, for example, extracting a migration destination column set in which the calculation result of the integration result has the smallest value.
At this time, the form of output is not limited, and output to a storage area, file output, screen output, and the like are assumed.

  As described above, in the present embodiment, by performing calculation for each migration source / destination in a plurality of column sets on a two-dimensional table, both results can be compared. A one-to-many relationship within one database can be easily grasped.

As described above, in the present embodiment,
For one column in the first table of interest in a two-dimensional table set that exists in multiple numbers,
A delimiter that divides the data in the column at a specifiable location,
Correlation rule calculation means for calculating a correlation rule as a correlation value between the divided data;
Correlation difference value calculating means for calculating a difference value between two combinations of the correlation value set;
The correlation rule calculation between the data held in the column is performed for the column set in another second table of interest, and further the correlation difference value calculation is performed between the two combinations of the correlation value set. A comparison calculation means for comparing the value calculation results with the results in the first table;
A table integration apparatus has been described that includes the above-described comparison result and a correspondence determination unit between a plurality of columns that determines whether or not the column of interest corresponds to a plurality of columns.

In the present embodiment,
The intra-table difference value calculation has the same combination as the data set existing in the table 1, the data combination existing in the table 2 as a comparison target,
In addition, the table integration apparatus provided with the correlation difference value calculation means for realizing the difference value calculation with the data not present in Table 2 as the correlation value 0 has been described.

In the present embodiment,
As a method for comparing the results of intra-table correlation difference value calculation between Table 1 and Table 2, the difference between the other correlation difference values having the same data set is taken from one correlation difference value set, and the correlation difference is calculated. Take the absolute value of the difference between the values,
The table integration apparatus provided with the comparison calculation means that uses the value (correlation comparison intermediate result) obtained by dividing the data set of Table 1 and the data set of Table 2 divided by the matched number has been described.

In the present embodiment,
Using the intermediate results of correlation comparison between Table 1 and Table 2, the support and confidence given as correlation values are given by the following formula using a variable α that can be specified.
α × (Intermediate comparison result of support level) + (1−α) × (Intermediate comparison result of confidence level)
The table integration device provided with the comparison calculation means used for determining the correspondence between the plurality of column sets has been described.

In the present embodiment,
Delimiter dividing means for determining whether or not to correspond to the contents obtained by dividing the plurality of column data existing in Table 1 with the column set in Table 2 as an input, a correlation rule calculating means, a correlation difference value calculating means, The table integration apparatus having the comparison calculation means and the determination means has been described.

In the present embodiment,
For one column in the first table of interest in a two-dimensional table set that exists in multiple numbers,
A delimiter step that divides the data in the column at a specifiable location,
A correlation rule calculation step of calculating a correlation rule as a correlation value between the divided data;
A correlation difference value calculating step for calculating a difference value between two combinations of the correlation value set;
The correlation rule calculation between the data held in the column is performed for the column set in another second table of interest, and further the correlation difference value calculation is performed between the two combinations of the correlation value set. A comparison calculation step for comparing the value calculation results with the results in the first table;
The table integration method including the above-described comparison result and the correspondence determination step between the plurality of columns for determining whether the target column corresponds to the plurality of columns has been described.

In the present embodiment, a case has been described in which the analysis target column of the migration source data is divided into two.
However, if the column to be analyzed in the migration source data has the same configuration as the migration destination column, for example, both the migration source data and the migration destination data have a configuration of “last name” column and “first name” column. If so, there is no need to split the migration source data column.

Embodiment 2. FIG.
FIG. 9 shows a configuration example of the table integration device 1 according to the present embodiment.
In addition to the configuration shown in FIG. 1, the table integration apparatus 1 according to the present embodiment includes a threshold acquisition unit 31, a calculation target designation variable acquisition unit 32, a threshold holding unit 168, and a correlation difference calculation target designation as shown in FIG. It has a variable holding unit 169.
9, elements using the same reference numerals as those in FIG. 1 are the same as those described in the first embodiment, and a description thereof will be omitted.
9, illustration of the migration source system 501, the migration source database 502, the migration destination system 601, and the migration destination database 602 shown in FIG. 1 is omitted.

In the present embodiment, it is determined whether or not a plurality of column sets correspond to the comparison calculation result appearing in the result of the comparison calculation unit 14 by the variable m acquired by the threshold acquisition unit 31.
In the first embodiment, the determination unit 15 extracts the column pair having the smallest integration result as a column pair candidate corresponding to the column of the migration source data.
On the other hand, in the present embodiment, the value of the threshold value m is variable, and the threshold value acquisition unit 31 acquires the value of the threshold value m from a user, an application program, or the like.
Then, the determination unit 15 sequentially extracts column pairs of correspondence candidates that are equal to or less than the value of m from which the value of the integration result is acquired, from the integration result of the minimum value.

Further, in the present embodiment, the correlation calculation target designation variable k (k> 1) acquired by the calculation target specification variable acquisition unit 32 includes the contents up to the top k of the correlation table in the correlation difference value calculation. Calculate the correlation difference value for the target.
In the first embodiment, the difference values of the support level and the certainty factor are calculated for all combinations of rows in each of the transfer source data and the transfer destination data.
For example, if there are 100 rows in the correlation table (FIG. 8) for both the migration source data and the migration destination data, the difference values of the support level and the certainty factor are calculated for each row with each of the other 99 rows.
On the other hand, in the present embodiment, the calculation target designation variable acquisition unit 32 acquires the value of the variable k from the user, application program, or the like, and the correlation difference value calculation unit 13 corresponds to the acquired value of k. Difference calculation is performed on the number of lines.
For example, if k = 10, the difference calculation of the support level and the certainty factor is performed between each of the 2nd to 11th rows for the first row of the correlation table for both the migration source data and the migration destination data. Ten difference values are obtained for each of the support level and the certainty level.
Similarly, for the second row of the correlation table, the difference calculation of the support level and the certainty factor is performed with each of the third to twelfth rows, and 10 difference values are obtained for each of the support level and the certainty level. Is obtained.

  The specification of the threshold value m and the specification of the correlation calculation target specification variable k are assumed not only by the user, but also by a file, an argument from another module, or the like.

  As described above, in the present embodiment, the determination criterion can be set by setting the threshold value m, and the number of calculation objects is narrowed down by the variable k, so that the amount of calculation can be reduced. .

In this embodiment,
Described a table integration device including a comparison calculation means for narrowing candidates by a threshold m that can be specified for a calculation result using a variable α for a correlation comparison intermediate result for support and a correlation comparison intermediate result for certainty. .

In the present embodiment,
The table integration apparatus provided with the correlation difference value calculation means for performing the intra-table correlation difference value calculation for the combination of selecting two from the set up to the highest k correlation values using the specifiable variable k has been described.

Embodiment 3 FIG.
FIG. 11 shows a configuration example of the table integration device 1 according to the present embodiment.
In addition to the configuration shown in FIG. 1, the table integration device 1 according to the present embodiment has a schema information analysis unit 41 and a schema information analysis result holding unit 170 as shown in FIG.
The schema information analysis unit 41 narrows down the determination by using the schema information and considering the order of the column names with respect to the result of step S4 in the first embodiment.
That is, the correspondence candidate column pairs are extracted with reference to the column arrangement order in the migration destination data together with the integration result value for each column pair in the migration destination data.
In the present embodiment, together with the determination unit 15, the schema information analysis unit 41 is an example of a correspondence candidate extraction process execution unit.
In FIG. 11, elements using the same reference numerals as those in FIG. 1 are the same as those described in the first embodiment, and description thereof is omitted.
In FIG. 11, the migration source system 501, the migration source database 502, the migration destination system 601, and the migration destination database 602 shown in FIG.

In this embodiment, when dividing and dividing in step S1, the order information of the front and rear portions is held in the definition information holding unit 161.
That is, the definition information holding unit 161 holds that the first half is “last name” and the second half is “first name”.
FIG. 10 shows a specific example of the third embodiment.
For each of the “last name” portion and “first name” portion obtained by dividing the “administrator” column of the migration source data, it is necessary to determine the corresponding column from the columns A, B, and C of the migration destination data.
From the results up to step S4, it is known that the “last name” portion of the “administrator” column corresponds to the column B, but the correspondence to the “first name” portion of the “administrator” column has not been found. .
At this time, the schema information analysis unit 41 acquires the order relationship of the divided portions of the “manager” column from the definition information holding unit 161 and grasps the order relationship of “last name” and “first name”.
Further, the correspondence information is grasped by obtaining the order information of the migration destination data from the definition information holding unit 161 and comparing it.
In the example illustrated in FIG. 10, the correspondence between “last name” − “first name” is “column B” − “column C” based on information that “last name” appears first and “first name” appears second. Determine the correspondence.
A result obtained by the schema information analysis unit 41 is stored in the schema information analysis result holding unit 170 and is input to the determination unit 15.

  As described above, in the present embodiment, by using the column definition information, it is possible to narrow down targets among similar columns and to determine whether a plurality of column sets are supported.

As described above, in the present embodiment,
A set of correspondence candidates between a plurality of column sets output from the judging means is input,
Definition information acquisition means for acquiring column order information from database definition information;
Table integration provided with schema information analysis means for determining one corresponding candidate from the correspondence candidate set between the plurality of column sets or analyzing that there is no corresponding candidate using the order information of the same column The apparatus has been described.

Embodiment 4 FIG.
FIG. 11 shows a configuration example of the table integration device 1 according to the present embodiment.
As shown in FIG. 11, the table integration device 1 according to the present embodiment includes a single column data analysis unit 42 and a single column data analysis result holding unit 171.
In FIG. 11, elements using the same reference numerals as those in FIG. 1 are the same as those described in the first embodiment, and description thereof is omitted.
In FIG. 11, the migration source system 501, the migration source database 502, the migration destination system 601, and the migration destination database 602 shown in FIG.

The single column data analysis unit 42 narrows down the correspondence relationship between the contents of the migration source table and the content of the migration destination table using the existing instance analysis method, and stores it in the single column data analysis result holding unit 171.
In the present embodiment, the single column data analysis unit 42 together with the association rule calculation unit 12 is an example of a second appearance tendency analysis processing execution unit.

In the single column data analysis, values are calculated by a known instance analysis method for the two data divided by the delimiter dividing unit 11, and further, values are calculated by a known instance analysis method for each column of the migration destination table, Compare the two.
A specific example will be described with reference to FIG.
The single column data analysis unit 42 pays attention to the “last name” column divided by the delimiter with respect to the column of the migration source data.
Then, the single column data analysis unit 42 counts the number of appearances for each instance of the field value appearing in the “last name” column, the instance of Sato 10 times, the instance of Tanaka 9 times, and so on. Organize (sort) in descending order.
Subsequently, the number of occurrences is counted and sorted in the same manner for all the migration destination columns.
Finally, if the number of appearance instances of the migration source data and the migration destination data match a certain number or more for the top N, the correlation rule calculation unit 12 selects the corresponding migration destination data column as an analysis target.
In the example of FIG. 12, the column A and the column B of the migration destination data are to be analyzed, while the column C is not to be analyzed.
Thereafter, the association rule calculation unit 12 calculates the support level and the certainty factor of the migration destination data by combining the column “name” (not shown) with the column “A”, which is the column of “last name”. The support level and the certainty factor are calculated by combining the column of “name” (not shown) with the column B which is the column of No. 1.
Subsequent processing procedures are as described in the first embodiment, and a description thereof will be omitted.

  As described above, in the present embodiment, candidates are narrowed down by instance comparison in units of columns, which is effective in reducing the amount of calculation in correlation rule calculation / correlation difference value calculation.

In this embodiment,
Single column data analysis means for estimating the correspondence between the column of the table of interest and the column of another table by comparing the data contents;
The table integration apparatus provided with the correlation rule calculation means for using the above result as a correlation value calculation target in the correlation rule calculation has been described.

Embodiment 5 FIG.
FIG. 13 shows a configuration example of the table integration device 1 according to the present embodiment.
The table integration device 1 according to the present embodiment receives one column set as input data to be migrated, inputs a plurality of columns as migration source data, and combines the data contents of the migration source column set. In order to determine whether a thing exists in the migration destination, as shown in FIG. 13, it has a delimiter division unit 11b, a correlation rule calculation unit 12b, a correlation difference value calculation unit 13b, a comparison calculation unit 14b, and a determination unit 15b. In this part, it is a mechanism having a function of reading a plurality of column sets of the transfer source for one piece of information of the transfer destination.
That is, in the first embodiment, when the “name” column exists in the migration source data and the “last name” column and the “first name” column exist in the migration destination data, “ Divide the “Name” column into the “Last Name” column and the “First Name” column, and select the “Last Name” and “First Name” column candidates corresponding to the divided “Last Name” and “First Name” column destination data The content is to extract from.
On the other hand, in the fifth embodiment, the migration destination data includes a “name” column, and the migration source data includes a plurality of “last name” and “first name” columns. Split the “name” column of the data into the “last name” column and the “first name” column, and the pair of the “last name” column and the “first name” column corresponding to the divided “last name” column and “first name” column The content is to extract from the migration source data.

  Also in the table integration device 1 according to the present embodiment, the schema information analysis unit 41 and the single column data analysis unit 42 may be used simultaneously.

  As described above, in the present embodiment, by having a configuration for determining the correspondence relationship between a plurality of migration source columns and a migration destination column, a migration that retains the combined contents of the data for the plurality of migration source columns It becomes possible to judge the destination.

Finally, a hardware configuration example of the table integration device 1 shown in the first to fifth embodiments will be described.
FIG. 15 is a diagram illustrating an example of hardware resources of the table integration device 1 described in the first to fifth embodiments.
15 is merely an example of the hardware configuration of the table integration device 1, and the hardware configuration of the table integration device 1 is not limited to the configuration described in FIG. Also good.

In FIG. 15, the table integration apparatus 1 includes a CPU 911 (also referred to as a central processing unit, a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, and a processor) that executes a program.
The CPU 911 is connected to, for example, a ROM (Read Only Memory) 913, a RAM (Random Access Memory) 914, a communication board 915, a display device 901, a keyboard 902, a mouse 903, and a magnetic disk device 920 via a bus 912. Control hardware devices.
Further, the CPU 911 may be connected to an FDD 904 (Flexible Disk Drive), a compact disk device 905 (CDD), a printer device 906, and a scanner device 907. Further, instead of the magnetic disk device 920, a storage device such as an SSD (Solid State Drive), an optical disk device, or a memory card (registered trademark) read / write device may be used.
The RAM 914 is an example of a volatile memory. The storage media of the ROM 913, the FDD 904, the CDD 905, and the magnetic disk device 920 are an example of a nonvolatile memory. These are examples of the storage device.
The “storage area 16” described in the first to fifth embodiments is realized by the RAM 914, the magnetic disk device 920, and the like.
A communication board 915, a keyboard 902, a mouse 903, a scanner device 907, an FDD 904, and the like are examples of input devices.
The communication board 915, the display device 901, the printer device 906, and the like are examples of output devices.

  The communication board 915 is connected to a LAN (Local Area Network), the Internet, a WAN (Wide Area Network), a SAN (Storage Area Network), etc., for example.

The magnetic disk device 920 stores an operating system 921 (OS), a window system 922, a program group 923, and a file group 924.
The programs in the program group 923 are executed by the CPU 911 using the operating system 921 and the window system 922.

The RAM 914 temporarily stores at least part of the operating system 921 program and application programs to be executed by the CPU 911.
The RAM 914 stores various data necessary for processing by the CPU 911.

The ROM 913 stores a BIOS (Basic Input Output System) program, and the magnetic disk device 920 stores a boot program.
When the table integration device 1 is activated, the BIOS program in the ROM 913 and the boot program in the magnetic disk device 920 are executed, and the operating system 921 is activated by the BIOS program and the boot program.

  The program group 923 stores programs that execute the functions described as “˜unit” and “˜means” in the description of the first to fifth embodiments. The program is read and executed by the CPU 911.

In the file group 924, in the description of the first to fifth embodiments, “read”, “determined”, “determined”, “calculated”, “calculated”, and “comparison of” ”,“ Evaluation of ”,“ analysis of ”,“ update of ”,“ setting of ”,“ registration of ”,“ selection of ”,“ extraction of ”,“ input of ”, Information, data, signal values, variable values, and parameters indicating the results of the processing described as “output of” are stored as items of “˜file” and “˜database”.
The “˜file” and “˜database” are stored in a recording medium such as a disk or a memory.
Information, data, signal values, variable values, and parameters stored in a storage medium such as a disk or memory are read out to the main memory or cache memory by the CPU 911 via a read / write circuit.
The read information, data, signal value, variable value, and parameter are used for CPU operations such as extraction, search, reference, comparison, calculation, calculation, processing, editing, output, printing, and display.
Information, data, signal values, variable values, and parameters are stored in the main memory, registers, cache memory, and buffers during the CPU operations of extraction, search, reference, comparison, calculation, processing, editing, output, printing, and display. It is temporarily stored in a memory or the like.
The arrows in the flowcharts described in the first to fifth embodiments mainly indicate input / output of data and signals.
Data and signal values are recorded on a recording medium such as a memory of the RAM 914, a flexible disk of the FDD 904, a compact disk of the CDD 905, a magnetic disk of the magnetic disk device 920, other optical disks, a mini disk, and a DVD.
Data and signals are transmitted online via a bus 912, signal lines, cables, or other transmission media.

In addition, in the description of the first to fifth embodiments, what is described as “to part” and “to means” may be “to circuit”, “to device”, and “to device”. It may be “˜step”, “˜procedure”, “˜processing”.
That is, the data processing method according to the present invention can be realized by the steps, procedures, and processes shown in the flowcharts described in the first to fifth embodiments.
In addition, what is described as “˜unit” and “˜means” may be realized by firmware stored in the ROM 913.
Alternatively, it may be implemented only by software, or only by hardware such as elements, devices, substrates, and wirings, by a combination of software and hardware, or by a combination of firmware.
Firmware and software are stored as programs in a recording medium such as a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, and a DVD.
The program is read by the CPU 911 and executed by the CPU 911.
That is, the program causes the computer to function as “to part” and “to means” in the first to fifth embodiments. Alternatively, the procedures and methods of “to part” and “to means” in the first to fifth embodiments are executed by a computer.

As described above, the table integration device 1 shown in the first to fifth embodiments includes a CPU as a processing device, a memory as a storage device, a magnetic disk, a keyboard as an input device, a mouse, a communication board, a display device as an output device, and a communication device. A computer including a board or the like.
As described above, the functions indicated as “˜unit” and “˜means” are realized by using these processing devices, storage devices, input devices, and output devices.

  DESCRIPTION OF SYMBOLS 1 Table integration apparatus, 11 Division | segmentation division | segmentation part, 12 Correlation rule calculation part, 13 Correlation difference value calculation part, 14 Comparison calculation part, 15 Judgment part, 16 Storage area, 17 Definition information acquisition part, 18 Data acquisition part, 19 Separation information Acquisition unit, 20 Database connection unit, 21 Connection information holding unit, 22 User I / F, 31 Threshold acquisition unit, 32 Calculation target designation variable acquisition unit, 41 Schema information analysis unit, 42 Single column data analysis unit, 101 Database definition information , 102 instance data, 103 delimiter information, 161 definition information holding unit, 162 acquisition data holding unit, 163 delimiter character information holding unit, 164 divided data holding unit, 165 correlation rule calculation result holding unit, 166 correlation difference calculation result holding unit, 167 Comparison calculation result holding unit, 168 threshold holding unit, 169 correlation Difference calculation target designation variable holding unit, 170 schema information analysis result holding unit, 171 single column data analysis result holding unit, 501 migration source system, 502 migration source database, 601 migration destination system, 602 migration destination database.

Claims (14)

Select a column pair to be analyzed as the first column pair to be analyzed for two-dimensional first data that contains multiple fields and each field is divided into one of multiple columns. A column pair selection process execution unit for executing the column pair selection process to be performed;
A first appearance tendency analysis processing execution unit for executing a first appearance tendency analysis process for analyzing an appearance tendency of a concatenated field value obtained by concatenating field values of each column of the first analysis target column pair in units of rows;
For two-dimensional second data in which a plurality of fields are included and each field is divided into one of a plurality of columns, the column pair to be analyzed is set as 1 as the second analysis target column pair. Second appearance trend analysis processing for selecting the pair or more and analyzing the appearance tendency of the connected field value obtained by connecting the field values of each column of the second analysis target column pair in units of rows for each second analysis target column pair A second appearance tendency analysis processing execution unit for executing
The analysis result for the first analysis target column pair and the analysis result for each second analysis target column pair are analyzed, and the concatenated field value in the first analysis target column pair is determined for each second analysis target column pair. A data processing apparatus comprising: an approximation degree calculation process execution unit that executes an approximation degree calculation process for calculating an approximation degree with an appearance tendency. The first appearance tendency analysis processing execution unit
For each linked field value in the first analysis target column pair, the appearance frequency is calculated,
The second appearance tendency analysis processing execution unit
For each second analysis target column pair, the appearance frequency is calculated for each connected field value in the second analysis target column pair,
The approximation calculation processing execution unit
The second analysis is performed by analyzing the calculated value of the appearance frequency for each linked field value in the first analysis target column pair and the calculated value of the appearance frequency for each linked field value in each of the second analysis target column pair. The data processing apparatus according to claim 1, wherein for each target column pair, the degree of approximation with the appearance tendency of the connected field value in the first analysis target column pair is calculated. The first appearance tendency analysis processing execution unit
As the appearance frequency of the linked field value, for each linked field value in the first analysis target column pair, calculate at least one of the support level and the certainty level,
The second appearance tendency analysis processing execution unit
As the appearance frequency of the connected field value, at least one of the support level and the certainty factor is calculated for each connected field value in the second analysis target column pair for each second analysis target column pair. The data processing apparatus according to claim 2. The approximation calculation processing execution unit
Calculating a difference between calculated values between linked field values in the first analysis target column pair as a first primary difference value, obtaining a plurality of first primary difference values in the first analysis target column pair;
For each second analysis target column pair, the difference between the calculated values between the linked field values in the second analysis target column pair is calculated as a second primary difference value, Obtaining a second primary difference value;
For each second analysis target column pair, the difference between the first primary difference value and the second primary difference value calculated from the combination of the same concatenated field values is calculated as the secondary difference value, and the second Obtain the secondary difference value for each analysis target column pair,
The degree of approximation with the appearance tendency of the connected field value in the first analysis target column pair is calculated for each second analysis target column pair using the secondary difference value. The data processing apparatus described. The approximation calculation processing execution unit
When the linked field value existing in the first analysis target column pair does not exist in the second analysis target column pair, the calculated value of the linked field value in the second analysis target column pair is set to 0, and the second primary The data processing apparatus according to claim 4, wherein a difference value is calculated. The first appearance tendency analysis processing execution unit
As the appearance frequency of the connected field value, the support level and the certainty factor are calculated for each connected field value in the first analysis target column pair,
The second appearance tendency analysis processing execution unit
As the appearance frequency of the connected field value, for each of the second analysis target column pair, the support level and the certainty factor are calculated for each connected field value in the second analysis target column pair,
The approximation calculation processing execution unit
A difference in support level between linked field values in the first analysis target column pair is calculated as a first support level primary difference value, and a plurality of first support level primary difference values in the first analysis target column pair are calculated. Get
For each second analysis target column pair, a support level difference between linked field values in the second analysis target column pair is calculated as a second support level primary difference value, and for each second analysis target column pair Obtaining a plurality of second support degree primary difference values;
For each second analysis target column pair, the difference between the first support degree primary difference value and the second support degree primary difference value calculated from the same combination field value combination is used as the support degree secondary difference value. Calculate
A difference in certainty between linked field values in the first analysis target column pair is calculated as a first certainty degree primary difference value, and a plurality of first certainty degree primary difference values in the first analysis target column pair are calculated. Get
For each second analysis target column pair, a difference in certainty between the connected field values in the second analysis target column pair is calculated as a second certainty factor primary difference value, and for each second analysis target column pair Obtaining a plurality of second certainty factor primary difference values;
For each second analysis target column pair, the difference between the first certainty factor primary difference value and the second certainty factor primary difference value calculated from the combination of the same concatenated field values is used as the certainty factor secondary difference value. Calculate
Using the support degree secondary difference value and the certainty degree secondary difference value, for each second analysis target column pair, calculating the degree of approximation with the appearance tendency of the connected field value in the first analysis target column pair The data processing apparatus according to claim 4, wherein: The approximation calculation processing execution unit
In the first analysis target column pair, for each connected field value, k first primary difference values are calculated between k (k> 1) other connected field values,
5. The k second primary difference values between k other concatenated field values are calculated for each concatenated field value in the second analysis target column pair. 6. The data processing device according to any one of 6. The second appearance tendency analysis processing execution unit
The data processing apparatus according to any one of claims 1 to 7, wherein column pairs of all combinations among a plurality of column pairs included in the second data are selected as second analysis target column pairs. The second appearance tendency analysis processing execution unit
9. The field value in each column included in the second data is analyzed, and a specific number of column pairs is selected as the second analysis target column pair based on the analysis result. The data processing apparatus described in 1. The column pair selection processing execution unit
The specific column included in the first data is divided into two, and the two columns after the division are selected as the first analysis target column pair, according to claim 1, Data processing device. The data processing device further includes:
Based on the degree of approximation for each second analysis target column pair calculated by the approximation degree calculation processing execution unit, from among the candidates for the second analysis target column pair that have a relationship corresponding to the first analysis target column pair The data processing according to claim 1, further comprising: a correspondence candidate extraction process execution unit that executes a correspondence candidate extraction process that extracts a correspondence candidate column pair that has a degree of approximation less than or equal to a variable m that can be specified. apparatus. The correspondence candidate extraction processing execution unit
12. The data processing apparatus according to claim 11, wherein a corresponding candidate column pair is extracted by referring to an arrangement order of columns in the second data together with an approximation degree for each second analysis target column pair. For a two-dimensional first data that includes a plurality of fields and each field is classified into one of a plurality of columns, the computer selects a column pair to be analyzed as a first analysis target column. Column pair selection processing to select as a pair,
A first appearance tendency analysis process in which the computer analyzes an appearance tendency of a concatenated field value obtained by concatenating field values of each column of the first analysis target column pair in units of rows;
For the two-dimensional second data in which a plurality of fields are included and each field is classified into one of a plurality of columns, the computer selects a column pair to be analyzed as a second analysis target. Select one or more pairs as column pairs, and analyze the appearance tendency of concatenated field values obtained by connecting the field values of each column of the second analysis target column pair in units of rows for each second analysis target column pair. Appearance trend analysis processing,
The computer analyzes the analysis result for the first analysis target column pair and the analysis result for each second analysis target column pair, and for each second analysis target column pair, in the first analysis target column pair A data processing method comprising: an approximation calculation process for calculating an approximation with an appearance tendency of linked field values. Select a column pair to be analyzed as the first column pair to be analyzed for two-dimensional first data that contains multiple fields and each field is divided into one of multiple columns. Column pair selection processing to
A first appearance tendency analysis process for analyzing an appearance tendency of a concatenated field value obtained by concatenating field values of each column of the first analysis target column pair in units of rows;
For two-dimensional second data in which a plurality of fields are included and each field is divided into one of a plurality of columns, the column pair to be analyzed is set as 1 as the second analysis target column pair. Second appearance trend analysis processing for selecting the pair or more and analyzing the appearance tendency of the connected field value obtained by connecting the field values of each column of the second analysis target column pair in units of rows for each second analysis target column pair When,
The analysis result for the first analysis target column pair and the analysis result for each second analysis target column pair are analyzed, and the concatenated field value in the first analysis target column pair is determined for each second analysis target column pair. A program that causes a computer to execute an approximation calculation process for calculating an approximation with an appearance tendency.

Images