JP2008197976A - Connection information generation program and connection information generation method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide connection information for connecting three or more data base tables. <P>SOLUTION: When receiving the URL or paths of database tables 11, 12, 13 and so on selected as the object of connection from an operator (step S101), this database connection device 20 generates a corresponding table 71 in which connection information acquired by associating each of the field names of the database tables, the other comparison object field names and the similarity of the both field names is recorded (steps S104 to S113). <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a program and a method for generating connection information for connecting a plurality of database tables.

  As is well known, many companies manage information about products and information about customers in a database. And when this kind of company merges or makes business alliance, when trying to mutually use each other's database, the database table of those databases may be connected.

  When concatenating multiple database tables, it is necessary to find a field common to each database table. Here, if the records of each database table have fields with the same field name, it is easy to connect the database tables. However, for example, even though a field in which a customer name is actually recorded exists in each database table, the field name may differ depending on the database table, such as a user name or a registrant name. . For this reason, it is practically not easy to connect a plurality of database tables.

  A technique for searching for a field whose field names are substantially common to each other in the two database tables is disclosed in Patent Document 1.

JP 2005-063332 A

  However, in the technology according to Patent Document 1 described above, the link information that associates substantially common field names is link information for linking two database tables, and three or more database tables are stored. There is no technique for generating connection information for connection.

  The present invention has been made in view of the above-described conventional circumstances, and an object of the present invention is to be able to generate connection information for connecting three or more database tables.

  A concatenated information generation program devised to solve the above problems identifies a computer selected from a plurality of database tables storing a plurality of records having a plurality of fields. A selection accepting unit that accepts selection information to be received from an operator through an input device, and a set of words that characterize each field name of the database table for each of the database tables specified by the selection information accepted by the selection accepting unit. Feature word set generation means for generating a feature word set based on the value of each field, feature vector generation means for generating a feature vector for each of all feature word sets generated by the feature word set generation means, selection Depending on the selection information received by the receiving means Processing for calculating the similarity of the remaining field names to be compared with the field names based on the feature vectors generated by the feature vector generation unit for all the field names of all the database tables to be processed; and , Similarity generation means for performing processing of associating and recording the comparison target field name and similarity in the similarity table corresponding to the field name, and all database tables specified by the selection information received by the reception means For each of all the field names, function as a recording means for recording, in the correspondence table in the storage device, linked information that associates the field name, the other field name to be compared, and the similarity of both field names. It is characterized by making it.

  With this configuration, when the operator selects several database tables as the connection target, the computer selects the field name and other comparison target field names for each of the database table field names. A correspondence table in which concatenation information in which the similarity between both field names is associated is recorded is generated. Since the concatenation information recorded in this correspondence table includes the similarity of both field names, the operator must grasp the field names having high similarity as the substantially common field names. Can do.

  Therefore, according to the present invention, connection information for connecting three or more database tables can be generated.

  The best mode for carrying out the present invention will be described below with reference to the accompanying drawings.

  FIG. 1 is a configuration diagram of a computer network system including three or more database devices 10 and a database connection device 20 of the present embodiment.

  Each of the database devices 10 is a general-purpose computer to which a database function is added, and is communicably connected to the database connection device 20 via a network. Accordingly, these database devices 10 include at least a storage, a CPU (Central Processing Unit), a memory, and a communication adapter, which are not shown. The storage is a storage device for storing programs and data. The CPU is a processing device that executes processing according to a program in the storage. The memory is a storage device in which programs and data read by the CPU are cached and a work area of the CPU is expanded. The communication adapter is a communication device for exchanging data with other computers. In order to realize the database function, any database device 10 searches the database table using the database tables 11, 12, 13,... In which records are stored in the storage and given search conditions. A program for storing the program.

  FIG. 2 is a configuration diagram of the database connecting device 20.

  The database connection device 20 is a personal computer to which a database connection function is added. Therefore, the database connection device 20 includes a display device 20a such as a liquid crystal display, an input device 20b such as a keyboard and a mouse, and a main body connected to these devices 20a and 20b. The main body includes at least a storage 20c, a CPU 20d, a memory 20e, and a communication adapter 20f. The database connection device 20 stores a database connection tool 21 in the storage 20c in order to realize a database connection function. The database connection tool 21 is a program for generating connection information for connecting a plurality of databases. The database connection tool 21 is executed by the CPU 20d that has received an execution instruction from the operator through the input device 20b.

  3 and 4 are diagrams showing a flow of processing by the database connection tool 21. FIG.

  After starting the execution of the database connection tool 21, in the first step S101, the CPU 20d accepts designation of the database tables 11, 12, 13,... To be connected from the operator through the input device 20b. Specifically, the CPU 20d displays a screen including an input field for inputting a URL [Uniform Resource Locator] and a path of the database tables 11, 12, 13,... On the display device 20a, and is input to the input field. By acquiring the URL and path, the specification of the database tables 11, 12, 13,... Selected by the operator as the connection target is accepted.

  The CPU 20d that executes this step S101 corresponds to the above-described selection receiving means.

  In the next step S102, the CPU 20d sets the extraction conditions for extracting the same or similar field names from the field names of all the database tables 11, 12, 13,... Specified as the connection target in step S101. And received from the operator through the input device 20b. The extraction conditions accepted here include “similarity threshold” and “rank threshold”. Among these, the “similarity threshold” is a lower limit value of similarity (described later) between field names that should be determined to be similar to each other. "Rank threshold" is the name of all fields other than the processing target field name when a certain field name is set as the processing target among the names of all the fields of all the database tables specified as the connection target. Is the lower limit value of the order of field names that should be determined to be similar to the processing target field name when they are arranged in descending order of similarity to the processing target field name. The CPU 20d displays a screen including input fields for inputting these “similarity threshold value” and “rank threshold value” on the display device 20a, and acquires the similarity threshold value and the rank threshold value input in the input field. The extraction condition is received from the operator.

  The CPU 20d executing step S102 corresponds to a similarity threshold receiving unit and a rank threshold receiving unit.

  In the next step S103, the CPU 20d is the same as or similar to the processing target field name when a certain field name is set as the processing target among all the field names of all the database tables designated as the connection target. Whether or not to include a field name in the same database table as the processing target field name in the comparison target to be determined is accepted from the operator through the input device 20b. Specifically, the CPU 20d displays a screen including an input field for inputting designation information for designating whether or not field names in the same database table are to be compared on the display device 20a. Through the input designation information, designation is made as to whether or not field names in the same database table are to be compared.

  The CPU 20d executing step S103 corresponds to a designation receiving unit.

  After step S103, the CPU 20d executes a first processing loop L1. In the first processing loop L1, the CPU 20d sequentially executes step S104 for each of all database tables specified by the designation received in step S101.

  In step S104, the CPU 20d executes objectification of the processing target database table. Here, the objectification of the database table will be described.

  FIG. 5 is a diagram illustrating an example of the data structure of the database table 11. 6 is a diagram showing an example of the structure of the objects 31, 34 and the tables 32, 33, 35 generated from the database table 11 of FIG.

  As shown in FIG. 5, each record of the database table 11 has a plurality of fields whose field names are employee number, name, address,. The CPU 20d generates a base class object 31, a property name list table 32, an instance list table 33, a plurality of instance objects 34, and a plurality of property value list tables 35 from the database table 11.

  The base class object 31 has the table name of the database table 11 as a class name. The base class object 31 has a pointer of the property name list table 32 and a pointer of the instance list table 33.

  The property name list table 32 has the same number of records as the fields of the records in the database table 11. Each record includes a “number” field in which a number for uniquely identifying the name of each field included in the record in the database table 11 is recorded, and a “property name” field in which the field name is recorded as a property name. have.

  The instance list table 33 has the same number of records as the database table 11. Each record has a “number” field in which a number for uniquely identifying each instance object 34 is recorded, and an “instance” field in which a pointer of the instance object 34 is recorded.

  The instance object 34 has an instance ID for uniquely identifying the instance and a pointer of the property value list table 35. One instance object 34 and one property value list table 35 are generated for each record in the database table 11. Accordingly, the same number of instance objects 34 and property value list tables 35 as the records of the database table 11 are generated.

  The property value list table 35 has the same number of records as the number of fields of one record of the database table 11. Each record has a field in which a number for uniquely specifying the value of each field in the corresponding record in the database table 11 is recorded, and a field in which the value of each field in the corresponding record is recorded. ing.

  Note that the objects 31, 34 and the tables 32, 33, 35 generated from the database table to be processed constitute one information system.

  The CPU 20d generates an object as shown in FIG. 6 for each of all the database tables specified by the designation received in step S101, then leaves the first processing loop L1 and proceeds to step S105 in FIG. To proceed.

  In step S105, the CPU 20d generates an integrated class object.

  FIG. 7 is a diagram illustrating an example of the structure of the integrated class object 41.

  As shown in FIG. 7, the integrated class object 41 has a class name, but does not have a pointer of an instance list table and a pointer of a property name list table. This class name is a name of a comprehensive information system formed by connecting all the information systems generated in step S104. After generating the integrated class object 41, the CPU 20d assigns the class name of the integrated class object 41 to each of the base class objects 31 as the class name of the upper class of each base class object 31 generated in step S104. Store. Note that the information systems are not connected to each other simply by associating the information systems with the integrated class object 41 in this way. By generating a correspondence table 71 described later, the information systems are connected to each other.

  In the next step S106 (FIG. 3), the CPU 20d generates a dependency information management object. The dependency information object is an object for managing the relationship between each information system before being associated with each other (their bases are each database table 11, 12, 13,...) And the overall information system after being associated with each other. It is.

  FIG. 8 is a diagram illustrating an example of the structure of the dependency information management object 51.

  As shown in FIG. 8, the dependency information management object 51 has the same class name as the class name of the integrated class object 41 of FIG. 7 as the name of the comprehensive information system. The dependency information management object 51 has a pointer of the pre-concatenation information system name list table 52. The pre-concatenation information system name list table 52 has the same number of records as all the database tables specified by the designation received in step S101. Each record includes a “number” field in which a number for uniquely identifying each database table is recorded, and a “pre-concatenation information system” in which a pointer of the pre-concatenation information system object 53 for managing each database table is recorded. ”Field. The pre-concatenation information system object 53 has the table name of each database table (that is, the class name of the base class object 31) as a class name and a pointer to the base class object 31.

  In the next step S107 (FIG. 3), the CPU 20d determines all the property names in the information systems associated with each other in step S105 (that is, all field names of all database tables specified by the designation received in step S101). ) For a feature word set. A feature word set refers to a set of words that characterize a word. That is, the CPU 20d extracts a phrase characterizing the field name from a value of the field in each record for a certain field name in a certain database table, and the extracted phrase is set as a characteristic word set in FIG. Are stored in association with the corresponding property names in the property name list table 32. FIG. 9 shows feature word sets generated for the “name” field, the “name” field, and the “employee name” field names, respectively. Various methods for generating a feature word set have been devised and disclosed, and will not be described here.

  The CPU 20d that executes step S107 corresponds to the feature word set generation unit described above.

  In the next step S108 (FIG. 3), the CPU 20d determines all the property names in each information system associated with each other in step S105 (that is, all field names of all database tables specified by the designation received in step S101). ), A feature vector is generated based on a feature word set associated with the property name. Various methods for generating a feature vector from a feature word set have been devised and disclosed, and will not be described here.

  The CPU 20d that executes this step S108 corresponds to the above-described feature vector generation means.

  After completing the generation of the feature word set and the feature vector, the CPU 20d executes the second processing loop L2. In the second processing loop L2, the CPU 20d executes step S109 one by one for each of all the property names (all field names) in all the property name list tables generated in step S104. .

  In step S109, the CPU 20d generates a processing target property object and a similarity table for the processing target property name.

  FIG. 10 is a diagram illustrating an example of the data structure of the processing target property object 61 and the similarity table 62.

  The processing target property object 61 is an object for associating the processing target property name with the similarity table 62 related thereto. This processing target property object 61 has a processing target property name and a pointer of the similarity table 62 related to the property name.

  The similarity table 62 is a table for recording similarities of all property names other than the processing target property name with respect to the processing target property name. The similarity table 62 has the same number of records as all the property names except the processing target property name. Each record has fields of “similar property name”, “similarity”, and “rank”. The “similar property name” field is a field in which the property name to be compared with the processing target property name is recorded. The “similarity” field is a field in which the distance between the feature vector of the property name and the feature vector of the processing target property name is recorded as the similarity. The “order” field is a field in which the order of the property names when all the property names other than the property names to be processed are arranged in descending order of similarity is recorded. Note that a method for obtaining a similarity between information elements based on a distance between feature vectors of a pair of information elements is disclosed in JP-A-2005-63332.

  The CPU 20d processes each property name in all the property name list tables 32 generated in step S104 (that is, all field names in all database tables specified by the designation received in step S101). After the target property object 61 and the similarity table 62 are generated, the process leaves the second processing loop L2, and the process proceeds to step S110 in FIG.

  In step S110, the CPU 20d determines whether or not the designation received in step S103 indicates that the property name (field name) of the same information system (database table) as the processing target property name is included in the comparison target. Determine. If the designation received in step S103 indicates that the property name having the same information system as the processing target property name is included in the comparison target, the CPU 20d branches the processing from step S110 to step S112. On the other hand, when the designation received in step S103 does not indicate that the property name having the same information system as the processing target property name is included in the comparison target, the CPU 20d advances the process to step S111.

  In step S111, the CPU 20d stores all the property names in all the property name list tables 32 generated in step S104 (that is, all field names of all database tables specified by the designation received in step S101). For each, the property name record included in the same information system is deleted from the property name similarity table 62. Whether a property name and a property name to be compared with it are included in the same information system includes two property names in all the property name list tables 32 generated in step S104. The determination is made based on whether or not the property name list table 32 exists.

  The CPU 20d performs processing for deleting records including property names of the same information system from the similarity table 62 for each of the property names, and then proceeds to step S112.

  In step S112, the CPU 20d removes records that do not match the extraction conditions (similarity threshold and rank threshold) received in step S102 from the similarity table corresponding to all property names. For example, when the similarity threshold value and the ranking threshold value are set to 0.7 and 3 by the operator, a record with a similarity of less than 0.7 or a ranking of 4th or less is obtained from each similarity table 62. Will be removed.

  The CPU 20d that executes these steps S109 to S112 corresponds to the above-described similarity generation unit.

  In the next step S113, the CPU 20d generates the correspondence table 71. The correspondence table 71 is a table in which all similarity tables 62 are integrated.

  FIG. 11 is a diagram illustrating an example of the data structure of the correspondence table 71.

  As shown in FIG. 11, each record of the correspondence table 71 includes fields of “number”, “table name 1”, “field name 1”, “similarity”, “field name 2”, and “table name 2”. have. The “number” field is a field in which a number for uniquely identifying each record is recorded. In the “table name 1” field, the class name of the information system (database table) including the property name (field name) associated with the similarity table 62 (the table name that the base class object 31 has as the class name) is recorded. Field. The “field name 1” field is a field in which a property name (field name) associated with the similarity table 62 is recorded. The “field name 2” and “similarity” fields are values in the “similar property name” and “similarity” fields in the similarity table 62 associated with the property names in the “field name 1” field. (Property name and similarity) are fields recorded respectively. The “table name 2” field is a field in which the class name of the information system (database table) including the value (property name) in the “field name 2” field (the table name that the base class object 31 has as the class name) is recorded. It is.

  The CPU 20d generates the correspondence table in FIG. 11 based on all the similarity degree tables 62, and then ends the processes according to FIGS.

  The CPU 20d that executes this step S113 corresponds to the recording means described above.

  Next, the operation and effect of the database connection device 20 of this embodiment will be described.

  When the operator executes the database connection tool 21 through the input device 20b, a screen (not shown) including input fields for inputting the URLs and paths of the database tables 11, 12, 13,... Is displayed on the display device 20a. It is displayed (step S101). The operator selects some arbitrary database tables, inputs their URLs and paths in the input fields, and clicks a decision button (not shown). Then, a screen (not shown) including an input field for inputting the extraction condition is displayed on the display device 20a (step S102). The operator inputs the values selected as the similarity threshold and the rank threshold in the input fields, and clicks a decision button (not shown). Then, a screen (not shown) including an input field for inputting designation information for designating whether or not to compare field names included in the same database table is displayed on the display device 20a (step S103). ). The operator inputs designation information in the input field and clicks a decision button (not shown). Then, the correspondence table 71 for linking several database tables selected by the operator is generated by the database linking device 20 (steps S104 to S113).

  In the generated correspondence table 71, a record in which the field name, the field name to be compared, and the similarity of both field names are associated is recorded as the connection information. Since this connection information includes the similarity between both field names, the operator can recognize field names having high similarity as substantially common field names.

  If the similarity threshold is set, only the link information having the similarity threshold or higher is recorded in the correspondence table 71 (step S109). Therefore, if the operator sets the similarity threshold as appropriate, the operator can remove the connection information having the similarity less than the threshold from the correspondence table 71 in advance.

  If a rank threshold is set, only the link information that is equal to or higher than the rank threshold in each similarity table 62 is recorded in the correspondence table 71 (step S109). Therefore, the operator can remove beforehand the connection information whose rank in each similarity table 62 is less than the threshold from the correspondence table 71 by appropriately setting the rank degree threshold.

  Furthermore, when it is specified not to compare field names included in the same database table, each similarity table 62 belongs to the same database table as the field name corresponding to the similarity table 62. The record of the comparison target field name is removed (step S110; YES, S111). Therefore, if the operator performs such designation in advance, the connection information of field names belonging to the same database table can be removed from the correspondence table 71 in advance.

  On the other hand, when it is specified to compare field names included in the same database table, each similarity table 62 belongs to the same database table as the field name corresponding to the similarity table 62. The record of the comparison target field name remains (step S110; NO). Therefore, if the operator performs such designation in advance, the association table 71 can also extract connection information between field names belonging to the same database table.

(Appendix 1)
Computer
Selection accepting means for accepting selection information for identifying some database tables selected from among a plurality of database tables storing a plurality of records having several fields from an operator through an input device;
A feature word set, which is a set of words that characterize each field name of the database table, is generated based on the value of each field for each of all database tables specified by the selection information received by the selection receiving unit. Word set generation means,
Feature vector generation means for generating a feature vector for each of all feature word sets generated by the feature word set generation means;
For each of all field names of all database tables specified by the selection information received by the selection receiving means, the feature vector generation means indicates the similarity of the remaining comparison target field names with respect to the field name. A similarity generation means for performing a process of calculating based on the generated feature vector, and a process of associating and recording the comparison target field name and the similarity in the similarity table corresponding to the field name;
For each of all the field names of all the database tables specified by the selection information received by the selection receiving means, the field name, the other comparison target field name, and the similarity of both field names are associated with each other A linked information generation program that causes linked information to function as a recording unit that records the linked information in a correspondence table in a storage device.

(Appendix 2)
Said computer further
The similarity threshold information specifying the lower limit value of the similarity is caused to function as a similarity threshold receiving unit that receives from the operator through the input device,
The similarity generation unit generates a similarity table for each of all field names, and then, from the similarity table, the similarity lower limit specified by the similarity threshold information received by the similarity threshold reception unit The connected information generation program according to appendix 1, wherein records that are less than the value are deleted.

(Appendix 3)
Said computer further
In the respective similarity tables, when the records are rearranged in the order of the degree of similarity, the rank threshold value information for specifying the lower limit value of the order is functioned as rank threshold value receiving means for receiving from the operator through the input device. ,
The similarity generation unit generates a similarity table for each of all field names, and then rearranges the records in the similarity table in order of similarity, and the rank threshold reception unit The connected information generation program according to appendix 1 or 2, wherein a record that falls below a lower limit value of the rank specified by the received rank threshold information is deleted.

(Appendix 4)
Said computer further
Designation information that designates whether or not the comparison target field name that belongs to the same database table as the field name corresponding to the similarity table should be recorded in the similarity table functions as a designation receiving unit that receives from the operator through the input device With
The similarity generation means specifies that the designation information received by the designation acceptance means does not record a comparison target field name belonging to the same database table as the field name corresponding to the similarity table in the similarity table If the similarity table is created for each of all the field names of all the database tables specified by the selection information received by the selection receiving means, the field names corresponding to the similarity table are different. The connected information generation program according to appendix 1, 2, or 3, wherein only the comparison target field name and similarity belonging to the database table are recorded in the similarity table.

(Appendix 5)
Computer
A selection acceptance procedure for accepting selection information for identifying several database tables selected from a plurality of database tables storing a plurality of records having several fields from an operator through an input device;
A feature word set, which is a set of words that characterize each field name of the database table, is generated based on the value of each field for each of all database tables specified by the selection information received in the selection receiving procedure. Word set generation procedure,
A feature vector generation procedure for generating a feature vector for each of all feature word sets generated in the feature word set generation procedure;
For each of all field names of all the database tables specified by the selection information received in the selection reception procedure, the similarity of the remaining comparison target field names with respect to the field name is determined by the feature vector generation procedure. A similarity generation procedure for performing processing to calculate based on the generated feature vector, and processing to record the comparison target field name and similarity in association with the similarity table corresponding to the field name; and
For each field name of all the database tables specified by the selection information received in the selection reception procedure, the field name, the other comparison target field name, and the similarity of both field names are associated with each other. A method for generating linked information, comprising: performing a recording procedure for recording linked information in a correspondence table in a storage device.

(Appendix 6)
The computer further comprises:
While executing a similarity threshold acceptance procedure for accepting similarity threshold information specifying the lower limit of the similarity through an input device from an operator,
In the similarity generation procedure, after generating a similarity table for each of all field names, a lower limit of the similarity specified by the similarity threshold information received by the similarity threshold receiving unit from the similarity table The linked information generation method according to appendix 5, wherein records lower than the value are deleted.

(Appendix 7)
The computer further comprises:
In the respective similarity degree tables, when the records are rearranged in the order of the similarity degree, the order threshold value receiving procedure for receiving the order threshold value information specifying the lower limit value of the order from the operator through the input device is executed. ,
In the similarity generation procedure, after generating a similarity table for each of all field names, the records in the similarity table are rearranged in the order of similarity, and further in the rank threshold value acceptance procedure. The linked information generation method according to appendix 5 or 6, wherein a record that falls below a lower limit value of the rank specified by the received rank threshold information is deleted.

(Appendix 8)
The computer further comprises:
Executes a designation acceptance procedure for accepting designation information for designating whether or not a comparison target field name belonging to the same database table as the field name corresponding to the similarity degree table should be recorded in the similarity degree table from the operator through the input device With
In the similarity generation procedure, the designation information received in the designation reception procedure specifies that a comparison target field name belonging to the same database table as the field name corresponding to the similarity table is not recorded in the similarity table If the similarity table is created for each of all the field names of all the database tables specified by the selection information received in the selection receiving procedure, the field name corresponding to the similarity table is different. The linked information generation method according to appendix 5, 6 or 7, wherein only the comparison target field name and similarity that belong to the database table are recorded in the similarity table.

(Appendix 9)
A selection receiving unit for receiving selection information for identifying several database tables selected from among a plurality of database tables storing a plurality of records having several fields from an operator through an input device;
A feature word set that is a set of words that characterize each field name of the database table is generated based on the value of each field for each of all database tables specified by the selection information received by the selection receiving unit. Word set generator,
A feature vector generation unit that generates a feature vector for each of all the feature word sets generated by the feature word set generation unit;
For each of all field names of all database tables specified by the selection information received by the selection receiving unit, the feature vector generation unit indicates the similarity of the remaining comparison target field names with respect to the field name. A similarity generation unit that performs a process of calculating based on the generated feature vector, and a process of recording the comparison target field name and the similarity in association with the similarity table corresponding to the field name; and
For each of all field names of all database tables specified by the selection information received by the selection receiving unit, the field name, the other comparison target field name, and the similarity of both field names are associated with each other. A connection information generation device comprising a recording unit that records connection information in a correspondence table in a storage device.

(Appendix 10)
A similarity threshold receiving unit that receives similarity threshold information specifying the lower limit of the similarity from the operator through the input device,
The similarity generation unit generates a similarity table for each of all field names, and then, from the similarity table, the similarity lower limit specified by the similarity threshold information received by the similarity threshold reception unit The connected information generating apparatus according to appendix 9, wherein records below the value are deleted.

(Appendix 11)
A rank threshold value receiving unit that receives rank threshold information for specifying a lower limit value of the order when the records are rearranged in order of the degree of similarity in each similarity table, from the operator through the input device; ,
The similarity generation unit generates a similarity table for each of all field names, and then rearranges the records in the similarity table in order of similarity, and the rank threshold acceptance unit The connected information generating apparatus according to appendix 9 or 10, wherein a record that falls below a lower limit value of the rank specified by the received rank threshold information is deleted.

(Appendix 12)
A designation receiving unit is further provided that receives designation information for designating whether or not the comparison target field name belonging to the same database table as the field name corresponding to the similarity table should be recorded in the similarity table from the operator through the input device. With
The similarity generation unit specifies that the designation information received by the designation receiving unit does not record a comparison target field name belonging to the same database table as the field name corresponding to the similarity table in the similarity table When the similarity table is created for each of all the field names of all the database tables specified by the selection information received by the selection receiving unit, the field name corresponding to the similarity table is different. The connected information generating apparatus according to appendix 9, 10 or 11, wherein only the comparison target field name and the similarity belonging to the database table are recorded in the similarity table.

Configuration diagram of a computer network system comprising three or more database devices and the database connection device of this embodiment Configuration diagram of database connection device Diagram showing the flow of processing by the database connection tool Diagram showing the flow of processing by the database connection tool Diagram showing an example of the data structure of the database table The figure which shows an example of the structure of the object and table which were produced | generated from the database table of FIG. Diagram showing an example of the structure of an integrated class object The figure which shows an example of the structure of a dependence information management object The figure which shows the feature word set which is respectively generated regarding the field name of “name” field, “name” field and “employee name” The figure which shows an example of the data structure of a process target property object and a similarity table The figure which shows an example of the data structure of a correspondence table

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Database apparatus 11 Database table 12 Database table 13 Database table 20 Database connection apparatus 20a Display apparatus 20b Input apparatus 20c Storage 20d CPU
21 Database connection tool

Claims (5)

Computer
Selection accepting means for accepting selection information for identifying some database tables selected from among a plurality of database tables storing a plurality of records having several fields from an operator through an input device;
A feature word set, which is a set of words that characterize each field name of the database table, is generated based on the value of each field for each of all database tables specified by the selection information received by the selection receiving unit. Word set generation means,
Feature vector generation means for generating a feature vector for each of all feature word sets generated by the feature word set generation means;
For each of all field names of all database tables specified by the selection information received by the selection receiving means, the feature vector generation means indicates the similarity of the remaining comparison target field names with respect to the field name. A similarity generation means for performing a process of calculating based on the generated feature vector, and a process of associating and recording the comparison target field name and the similarity in the similarity table corresponding to the field name;
For each of all field names of all the database tables specified by the selection information received by the receiving unit, the field name, the other comparison target field name, and the similarity of both field names are associated with each other A linked information generation program that causes information to function as a recording unit that records information in a correspondence table in a storage device. Said computer further
The similarity threshold information specifying the lower limit value of the similarity is caused to function as a similarity threshold receiving unit that receives from the operator through the input device,
The similarity generation unit generates a similarity table for each of all field names, and then, from the similarity table, the similarity lower limit specified by the similarity threshold information received by the similarity threshold reception unit 2. The linked information generation program according to claim 1, wherein records having a value lower than the value are deleted. Said computer further
In the respective similarity tables, when the records are rearranged in the order of the degree of similarity, the rank threshold value information for specifying the lower limit value of the order is functioned as rank threshold value receiving means for receiving from the operator through the input device. ,
The similarity generation unit generates a similarity table for each of all field names, and then rearranges the records in the similarity table in order of similarity, and the rank threshold reception unit The connected information generation program according to claim 1 or 2, wherein a record that falls below a lower limit value of the rank specified by the received rank threshold information is deleted. Said computer further
Designation information that designates whether or not the comparison target field name that belongs to the same database table as the field name corresponding to the similarity table should be recorded in the similarity table functions as a designation receiving unit that receives from the operator through the input device With
The similarity generation means specifies that the designation information received by the designation acceptance means does not record a comparison target field name belonging to the same database table as the field name corresponding to the similarity table in the similarity table If the similarity table is created for each of all the field names of all the database tables specified by the selection information received by the selection receiving means, the field names corresponding to the similarity table are different. 4. The linked information generating program according to claim 1, wherein only the comparison target field name and the similarity belonging to the database table are recorded in the similarity table. Computer
A selection acceptance procedure for accepting selection information for identifying several database tables selected from a plurality of database tables storing a plurality of records having several fields from an operator through an input device;
A feature word set, which is a set of words that characterize each field name of the database table, is generated based on the value of each field for each of all database tables specified by the selection information received in the selection receiving procedure. Word set generation procedure,
A feature vector generation procedure for generating a feature vector for each of all feature word sets generated in the feature word set generation procedure;
For each of all field names of all the database tables specified by the selection information received in the selection reception procedure, the similarity of the remaining comparison target field names with respect to the field name is determined by the feature vector generation procedure. A similarity generation procedure for performing processing to calculate based on the generated feature vector, and processing to record the comparison target field name and similarity in association with the similarity table corresponding to the field name; and
For each field name of all the database tables specified by the selection information received in the reception procedure, the field name, the other field name to be compared, and the similarity of both field names are associated with each other A method for generating linked information, comprising: executing a recording procedure for recording information in a correspondence table in a storage device.