JP2011248661A - Database control device, database control method, program and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a database control device, a database control method, a program and a recording medium that serve to enhance identity among data contents of information searched for among different databases and thereby to enhance the accuracy of information search among a plurality of databases by homogenizing data contents in the same item of information among different databases so that reuse is possible no matter which one of multiple databases may be referenced by whatever service or application.SOLUTION: A database control device has a similarity calculating unit that compares data contents of reference data accumulated in a reference database for reference use and of accumulation data accumulated in an accumulation database for accumulation use, figures out any similarity between the accumulation data and the reference data, and detects any of the accumulation data whose figured-out similarity surpasses a preset threshold as similar data, and a data contents updating unit that updates the accumulation data detected as the similar data with the data contents of the accumulation data.

Description

  The present invention relates to a database control device, a database control method, a program, and a recording medium.

A database is an information processing system that accumulates information flooded in the city in an advanced information society and provides information to users as needed.
For this reason, the proposition of how to manage information in order to facilitate data retrieval increases day by day due to the nature of storing all kinds of information, from important secrets to miscellaneous records, in the database. ing.

In addition, reusability of information stored in each of a plurality of databases managed as described above is also an important property to be discussed.
For example, the use of searching for information identical to information retrieved from a certain database from another database is considered. If the same information is not stored in the same configuration between the two databases, the target data cannot be extracted by a simple search process, and appropriate information is obtained from the database in a limited time. As a result, the convenience of the entire system is impaired.

Currently, there is no standard or specification that unifies data contents among multiple databases, and even if it is assumed that the standard or specification has been established, the database system generally records various records as described above. Since it is used in a storing configuration, it is considered that the possibility that such a standard or specification is widely used is extremely low.
For example, an information search method is disclosed in which a common information search language is transmitted to a plurality of databases having different configurations and methods, and information search is performed using the common information search language (see, for example, Patent Document 1). ).
In this information retrieval method, when an end user inputs an information retrieval procedure and an editing procedure from an information input / output device in a format in accordance with the specifications of a “common” information retrieval language, the generation means is the target data management system. A resource (control statement, source program, parameter, command, etc.) that can be searched is generated, and the corresponding execution means is activated internally. The activated execution means uses the resource provided by the generation means and performs a search process from the database. The search result is delivered to the sending means, and the result edited in accordance with the editing procedure is visually displayed on the information input / output device at hand of the end user.

In addition, a technique in which incompatible databases exchange data bidirectionally is disclosed (for example, see Patent Document 2).
In this exchange technique, the data set in the second database system is updated by updating the data set in the database system by computer processor instructions stored in the computer and by the changes made to the data set in the first database system. The set is updated. Thereby, the reusability of a data set can be improved in the first and second database systems.

JP-A 61-234431 JP-T-2003-511796

  However, although the information search method of Patent Document 1 ensures information reusability only in database search, the same data is obtained for the accuracy of information stored in each database itself, that is, the search key. Therefore, new development costs are always generated as a process for improving the homogeneity of information when diverted to searches other than the same search key.

  In addition, since the data exchange between the databases of the above-mentioned Patent Document 2 requires that processing be performed after a data set for mutual data exchange is designated in advance, a new database group can be renewed. Each time a database is added, there is a cost associated with the new database so that it can be reused in units of data sets.

  An object of the present invention is to provide a database control apparatus, a database control method, a program, and a recording medium that solve such problems.

  The present invention has been made to solve the above-described problems, and the database control apparatus according to the present invention includes reference data stored in a reference database for reference and stored data stored in a storage database for storage. The similarity calculation unit that detects the accumulated data in which the obtained similarity exceeds a preset threshold value as similar data, compares the data content with the reference data, obtains the similarity between the accumulated data and the reference data And a data content update unit for updating the stored data detected as the same data with the data content of the stored data.

  In the database control apparatus according to the present invention, each of the accumulated data and the reference data is composed of a plurality of pieces of configuration data, and availability information indicating whether or not to update the information for each configuration data is set in advance. The information processing apparatus further includes a rewrite item detection unit that determines, for each piece of configuration data, whether or not to update the accumulated data based on the data content of the reference data.

  When the availability information set in the configuration data is acceptable, the database control apparatus according to the present invention corresponds to the reference data corresponding to the configuration data of the stored data that is the same data as the reference data. On the other hand, if the availability information is NO, the configuration data of the stored data that is the same data as the reference data is changed by the configuration data corresponding to the reference data. It is characterized by not.

  When the availability information is acceptable, the database control device of the present invention overwrites the corresponding configuration data in the data of the reference database with the configuration data, or the configuration data of the data and the reference database A change type as to whether the data is a union with corresponding configuration data is set.

  In the database control apparatus according to the present invention, the similarity calculation unit is configured to calculate similarity between the configuration data of the reference data and the configuration data of the accumulated data, which is set for each data type of the configuration data. The similarity is calculated for each of the configuration data in the data of the reference database and the storage database by a degree calculation method, and the sum of the calculated similarities is used as the similarity of the data.

  In the database control apparatus of the present invention, the meta information included in the configuration data indicates a data type, and the similarity calculation unit includes a calculation method table indicating a relationship between the related data type and the similarity calculation method. And extracting meta information from the configuration data, selecting a similarity calculation method corresponding to the extracted meta information from the calculation method table, and calculating the similarity for each of the configuration data .

  In the database control apparatus of the present invention, the similarity calculation unit multiplies the similarity obtained by the similarity calculation method by a weighting coefficient set in the configuration data, and thereby calculates the similarity of the configuration data. It is characterized by.

  The database control device of the present invention includes a reliability data storage unit that stores reliability data indicating the reliability of a database composed of a plurality of items updated at regular intervals for each database, and the reliability data storage unit based on the reliability data. It further has a reference database setting unit that obtains the reliability for each database and sets the database with the highest reliability as the reference database.

  In the database control apparatus of the present invention, the reliability data includes at least the number of external links to the server managing the database, the presence / absence of access information, the number of accesses by the user, the number of site updates, and the advertisement display displayed on the site. It is any one of the area and the occupying ratio, or a combination thereof.

  The database control method of the present invention is a database control method for operating a database control apparatus that uniformizes the data content of stored data stored in each database among a plurality of databases. The data contents of the reference data stored in the reference database for storage and the stored data stored in the storage database for storage are compared, and the degree of similarity between the stored data and the reference data is obtained. Similarity calculation process for detecting the accumulated data whose similarity exceeds a preset threshold as similar data, and the data content update unit detects the accumulated data detected as the similar data as data of the accumulated data A data content updating process for updating the content.

  The program of the present invention is a program that causes a computer to execute the function of a database control device that equalizes the data content of accumulated data accumulated in each database among a plurality of databases. The data contents of the reference data stored in the reference database for storage and the stored data stored in the storage database for storage are compared, and the degree of similarity between the stored data and the reference data is obtained. Similarity calculation processing for detecting, as similar data, the accumulated data in which the similarity exceeds a preset threshold value, and the accumulated data detected by the data content update unit as the similar data are stored as data of the accumulated data. A computer-executable program having data content update processing to be updated according to content That.

  The recording medium on which the program of the present invention is recorded is a program that causes a computer to execute the function of a database control device that uniformizes the data content of stored data stored in each database among a plurality of databases. The calculation unit compares the data contents of the reference data stored in the reference database for reference and the stored data stored in the storage database for storage, and determines the similarity between the stored data and the reference data. A similarity calculation process for detecting the stored data in which the calculated similarity exceeds a preset threshold as similar data, and the data content update unit detects the stored data detected as the similar data, Executed by a computer having a data content update process that updates the data content of the stored data A recording medium recording the ability programs.

  The present invention makes it possible to easily maintain a highly reusable database of accumulated data.

It is a figure explaining the concept of the service system which consists of several service servers which have a database using the database control apparatus by this invention. It is a schematic block diagram which shows the structural example of the database management system which equalizes the data of the database using the analysis server 1 by this embodiment. It is a block diagram which shows the structural example of the analysis server 1 in FIG. 4 is a table showing an example of reference database data provided in a server A and storage database data provided in a server B. It is a flowchart which shows the operation example of the process which the analysis server 1 updates the accumulation | storage data of the data of a target database with the reference data of a reference database, and equalizes data. FIG. 5 is a diagram illustrating a process in which the data content update unit updates the content of stored data in the stored database in FIG. 4 with information of reference data in a reference database. It is a figure which shows the structural example of the reference | standard table in which the reliability reference | standard stored in the reliability data storage part 18 used for calculating | requiring the reliability of a database is shown. FIG. 8 shows that the reference database setting unit 15 selects one of a plurality of databases as the most reliable reference database based on each reliability standard of the standard table stored in the reliability data storage unit 18 shown in FIG. It is a flowchart which shows the operation example of whether to do. FIG. 9 is a flowchart illustrating an operation example in which the similarity calculation unit 16 calculates the similarity between the reference data in the reference database in FIG. 4 and the accumulated data in the target database. It is a figure which shows the data structure of reference data and the accumulation | storage data explaining the calculation process of the similarity degree of the reference data of a reference database, and the accumulation | storage data of a target database. It is a figure explaining the calculation process of the similarity for every item of data.

Hereinafter, the concept of the present invention will be described with reference to the drawings. FIG. 1 is a diagram for explaining the concept of a service system including a plurality of service servers having a database, using a database control apparatus according to the present invention. The service system in FIG. 1 is an information search system, for example, and includes an analysis server 100, a plurality of service servers 102 and 105, and a user terminal 107.
The analysis server 100 is a database control device in which an analysis module 101 that performs processing of homogenizing (that is, homogenizing) data contents between different databases is installed as an application.
Accordingly, the contents of the data 103 and the data 106 of the service server 102 and the service server 105 in which data of similar fields are accumulated are made uniform.

For example, the service server 105 has a higher access count from the user terminal 107 for searching for information and receiving the data search service 104 than the service server 102, and the content of the data 106 is compared to the data 103. Assume that it is set to be more substantial and the data accuracy is high.
In this case, the analysis server 100 compares the data 106 of the service server 105 and the data 103 of the service server 102 and detects a data set having high similarity.
Then, the analysis server 100 performs a process of updating the data 103 of the service server 102 with the data 106 of the service server 105, assuming that the sets with high similarity are the same type of data.

As a result, the contents of the data 103 of the service server 102 are made uniform with the data 103 of the service server 105 that is assumed to have high accuracy.
Therefore, the user terminal 107 that has accessed the service server 102 without knowing the existence of the service server 105 can also obtain the data 103 having the same accuracy as the data 106 of the service server 105. That is, since the data in each database of the service server group managed by the analysis server 100 is uniform, the user can obtain information with the same accuracy from the user terminal 107 regardless of which service server is accessed. Can do.

Hereinafter, an analysis server as a database control apparatus according to an embodiment of the present invention will be described in detail.
FIG. 2 is a schematic block diagram showing a configuration example of a database management system for uniformizing database data using the analysis server 1 according to the present embodiment. This analysis server 1 corresponds to the analysis server 100 of FIG.
Each of these service server 21 to service server 2n has databases DB1, DB2,..., DBn used for searching data in the same field.
The analysis server 1 performs data equalization processing between the databases DB1 to DBn at a predetermined cycle (for example, every day). The analysis server 1 and each of the service servers 21, 22, ..., 2n are connected by a network 200. The network 200 is configured by a data communication network such as the Internet.

FIG. 3 is a block diagram illustrating a configuration example of the analysis server 1 in FIG. In this figure, the analysis server 1 includes a control unit 11, a data input unit 12, a data transmission / reception unit 13, a data content update unit 14, a reference database setting unit 15, a similarity calculation unit 16, a rewrite item detection unit 17, a reliability data. A storage unit 18, a storage unit 19, and a display device 20 are provided.
The data input unit 12 takes in data (including a control instruction for the analysis server 1) input by a user from an input device (not shown).
The data transmission / reception unit 13 is under the control of the analysis server 1 and transmits / receives data and control signals to / from each service server having a database to be homogenized.
The reference database setting unit 15 uses any reference data (reliability data indicating the reliability of the database) indicating the reliability of each of the databases DB1 to DB2n provided in the reliability data storage unit 18. Then, it is determined whether or not the reference data is referred to when the data stored in the storage database for storage is rewritten. That is, the reference database setting unit 15 determines which database will be a reference database to be referred to when changing data in another database when the data in the database is uniformed, and according to the determination result. Select the reference database and set the other database as the target database. Here, the reference database is a database in which the data content of the accumulated data is enhanced compared to the accumulation database (target database) and the most reliable information is obtained from the databases managed by the database controller. Selected. The selected database is a reference database, and a database managed by a database controller other than the reference database is a storage database.

The similarity calculation unit 16 selects, for each of the reference data stored in the reference database, reference data for calculating the similarity one by one from a plurality of reference data, and each stored data stored in the target database The similarity is calculated. Then, the similarity calculation unit 16 calculates the similarity of the accumulated data (accumulated data of the target database) having the highest similarity and exceeding a preset threshold with respect to each of the reference data. It is extracted as similar data having the same type of information content as the reference data. Here, the similarity calculation unit 16 calculates the similarity used for determining whether or not the reference data and the accumulated data are the same type of data, and the reference item data constituting each of the reference data and the accumulated data. Depending on the data type of the stored item data, the similarity is calculated by several methods. For example, when the data is composed of character strings, the identification information indicated by numerical values is obtained by the ratio of the same word or part of the words included in each character string of the reference item data and the accumulated item data. In the case of, the value is calculated based on whether or not the numerical values are exactly the same, and if the numerical value is meaningful such as a price, the numerical value is calculated based on the proximity of the numerical values (the amount of difference between the numerical values) (details will be described later) ).
Here, the similarity calculation unit 16 indicates the data type described in the meta information included in each piece of configuration data, and the similarity indicating how the similarity is calculated corresponding to the data type. Has a calculation method table that shows the relationship with the calculation method (a table in which the similarity calculation method is set for each type of data), extracts meta information from the configuration data, and supports data types corresponding to the extracted meta information The similarity calculation method is selected from the calculation method table, and the similarity for each configuration data is calculated. The meta information refers to additional data having a high degree of abstraction about the data itself attached to the data.

The rewrite item detection unit 17 detects accumulated item data to be rewritten from accumulated item data constituting accumulated data extracted as the same data.
The data content updating unit 14 rewrites the rewrite accumulation item data for rewriting the detected accumulation data with the information of the accumulation item data of the corresponding reference data.
The storage unit 19 stores information on which database is the reference database and which database is the target database in the service server database managed by the analysis server 1, and the similarity used by the similarity calculation unit 16. The threshold value and the like are stored.
The display device 20 displays data input by the user, the progress of the homogenization process performed on each database, and the like. The control unit 11 controls the various operations described above.

FIG. 7 is a diagram showing a configuration example of a reference table showing reliability criteria stored in the reliability data storage unit 18 used for determining the reliability of the database. This reference table is provided for each database managed by the analysis server 1.
As the reliability standard, for example, legitimacy, organization information, expertise, related person, number of accesses, number of updates, advertisement area, etc. are used as the standard name. As variable names corresponding to these reference names, ie, legitimacy, organization information, expertise, related person, access count, update count, and advertisement area, “Validity”, “Organization”, “Speciality”, “Person”, respectively. , “Access”, “Update”, and “Advertisement” are set.

In the standard table of FIG. 7, as the contents of each standard, the legitimacy is based on the presence or absence of grounds for securing the site information. The number of links, that is, the number of other service servers in which the service server to be determined is set as the link destination.
Organization information is based on the presence or absence of an affiliated organization in the real world (whether or not affiliated with an existing organization), and since it is used to determine the criteria, the collected information is the existing organization. Information indicating whether or not access data (access information for contacting the organization such as the organization name, address, telephone number, and mail address) is described.
Expertise is based on the presence or absence of proof of expertise of the organization to which it belongs, and since it is used to determine the criteria, the information collected is the qualification issued by a third party or the acceptance criteria for the qualification. This is information indicating whether or not disclosure information is described.

The related person is based on the presence or absence of a link with a real person who actually exists in the real world, and information used to determine the reference includes access data (address, telephone). This is information indicating whether or not (access information for contacting a person such as a number or e-mail address) is described.
In addition, the number of accesses is based on the number of accesses from general users, and since it is used to determine the reference, the collected information includes the number of unique users accessing the service server search site and PV (page view, This is information obtained by quantifying access such as the number of page views.
In addition, the number of updates is based on the number of updates of the search site of the service server, and since it is used for criteria determination, the information collected is the number of updates updated since the search site was created, or It is the information which shows the update frequency which shows the frequency | count of an update in the period (for example, half year) set beforehand.

The advertising area is based on the ratio of the display area of advertisements (advertisements) published on the service server site to the display area of the entire site, and is used for criteria judgment. Is information indicating the display area of the advertisement included in the site, the display area of the entire site, and the ratio of the display area of the advertisement to the display area of the entire site.
A weighting coefficient is set for the above-described criteria. For example, in this embodiment, a weighting factor “5” is assigned to the legitimacy criterion, a weighting factor “5” is assigned to the organization information criterion, The weight is “10” for the criterion, the weight is “5” for the criterion for the related person, “30” for the criterion for the number of accesses, “30” for the criterion for the number of updates, and “15” for the criterion for the advertisement area. Is set.
In the reference table, for each reliability name of each variable name, a numerical value or T / F is used as a data type indicating whether the information to be collected is a numerical value or whether it corresponds to the reference. (True / False, true / false).
The information collected by each reliability standard in the above-described standard table is periodically collected by the operator through the data input unit 12 to check the obtained reliability. To the reference table of the reliability data storage unit 18.

Next, before explaining the operation of the present embodiment, a configuration example of data to be uniformized will be described with reference to FIG. In the following, for the sake of simplicity, the service servers under the management of the analysis server 1 in FIG. 2 are assumed to be two service servers 21 and 22, and the data configuration in the databases DB1 and DB2 will be described. In this embodiment, a service server having a reference database is set as server A, and a service server having a storage database (target database) is set as server B.
FIG. 4 is a table showing an example of reference database data provided in the server A and storage database data provided in the server B. In the reference database and the storage database, a plurality of such data is stored as stored data. As item name of reference item data constituting reference data or item name of accumulated item data constituting accumulated data, product name (variable name: item_name), comment (variable name: comment), price (variable name: price), Search key (variable name: keyword), JAN (Japan Article Number code, product identification number and bar code standard defined in Japanese Industrial Standards) code (variable name: jan_code), storage method (variable name: bin_kbn) The initial letter (variable name: Initial) of the product is described.

The product name is described as word information as information indicating that a plurality of words are connected by the symbol “#” as meta information of the item. For example, as the stored item data, “Company A # Bird Fried # 6 #Frozen” is described in the database DB1, and “Company A # Fried # 6 Chickens” is written in the database DB2.
In addition, a character string is described as information indicating that the comment is a sentence with a series of characters as meta information of the item. For example, as stored item data, “DB is soft even if it is cold, and it is succulent because it is juicy. "It's a juicy fried chicken. It is characterized by bouillon and mellow taste of domestic chicken."

A numerical value is described as information indicating that the price is a number as meta information of the item. For example, as the accumulated item data, “350” is described in the database DB1, and “348” is described in the database DB2.
The search key is described as meta information of the item, and word duplication is described as information indicating that a plurality of words for retrieving each stored data from the database are connected by a symbol “#”. For example, as the accumulated item data, “tori # bird # chicken # fried # fried # fried # fried # caraage # frozen food # cold food” is described in the database DB 1, and “A company # fried # from # "6" is described.

An ID (identify) indicating that the JAN code is product identification information is described as item meta information. For example, as the accumulated item data, “5511000044223333” is described in the database DB1, and is not described in the database DB2.
In the storage method, a single word indicating that it is indicated by one word as meta information of an item is described. For example, as the stored item data, “Frozen” is described in the database DB1, and “Frozen” is described in the database DB2.
A single word is described as information indicating that the initial of the product name is composed of one word. For example, as the stored item data, the initial “TO” is written in the database DB1, and is not described in the database DB2.

A weighting coefficient to be multiplied when the information of the accumulated item data of the reference data matches the accumulated item data of the accumulated data is set for each item.
In the product name, “30” is described as a weighting coefficient to be multiplied with respect to the number of matched words among a plurality of words composed of words. “5” is described as a weighting coefficient to be multiplied to the number of matching words among the words constituting the character string of the comment. In the price, “30” is described as a weighting coefficient for multiplying the numerical value obtained by dividing the smaller one by the larger numerical value. In the search key, “5” is described as a weighting coefficient to be multiplied to the number of matched words among a plurality of words constituted by word duplication. The JAN code describes a weight coefficient “10” that is multiplied by a numerical value given when the identification numbers of the codes match. The storage method describes a weighting factor “10” that is multiplied by a numerical value given when words described as single words match. The initial value of the product name describes a weighting factor “10” that is multiplied by a numerical value given when a word described as a single word matches.

Update information indicating whether or not to rewrite the stored item data of the stored data and how to process the rewritten data is set for each item based on the reference item stored item data. Here, whether or not rewriting is performed is permission information. When update information is non-target, it indicates that rewriting is not possible, and when update information is overwritten or union, it indicates that rewriting is possible. ing.
In the item of product name, overwriting is described as information indicating that the product name information of the accumulated data is rewritten with the product name of the reference data. In the comment section, non-object is described as information indicating that the comment information of the accumulated data is not rewritten by the comment of the reference data. In the price section, non-object is described as information indicating that the price information of the stored data is not rewritten according to the price of the reference data. The search key field describes a union that is information indicating that the OR of the reference data information and the stored data information is taken. In the JAN code section, overwriting is described as information indicating that the JAN code ID number of the stored data is rewritten with the JAN code ID number of the reference data. In the storage method section, overwriting is described as information indicating that the storage data storage method information is rewritten by the reference data storage method. In the section of the initial of the product name, overwriting is described as information indicating that the initial information of the product name of the stored data is rewritten by the initial of the product name of the reference data.

  Next, the operation of the database data equalization processing performed by the analysis server 1 in this embodiment will be described with reference to FIG. FIG. 5 is a flowchart showing an operation example of processing in which the analysis server 1 updates the data in the target database with the data in the reference database and uniformizes the data. Assuming that the service servers managed by the analysis server 1 are two service servers 21 and 22, the equalization processing in the analysis server 1 will be described below.

The data input unit 12 starts the homogenization process in the analysis server 1 when information indicating the execution of the homogenization process is input from an operator (user). Moreover, the control part 11 will start the equalization process in the analysis server 1, if the time used as the predetermined period set beforehand passes.
The reference database setting unit 15 determines whether one of the database DB1 of the service server 21 and the database DB2 of the service server 22 is highly reliable and is used as a reference database (step S1).
At this time, the reference database setting unit 15 sets the higher reliability as the reference database, sets the service server having the reference database in the server A, and sets the lower one as the storage database to be updated with the data of the reference database. A service server having a database is set in the server B, and the similarity calculation unit 16 is notified that the setting has been completed.

Next, the similarity calculation unit 16 obtains the number of data in the reference database provided in the server A by accessing the server A via the data transmission / reception unit 13 (step S2). That is, the analysis server 1 transmits to the server A a request signal for acquiring the number of data stored in the reference database (that is, the number of records) and the record number indicating the position where each data is stored. When this request signal is received via the data transmission / reception unit 13 and the network, the server A acquires the number of records, which is the number of reference data stored in the storage database, and the record number of each data. The number of reference data and the record number of the data are transmitted to the analysis server 1.
When the similarity calculation unit 16 receives the reference data number and record number of the data stored in the reference database from the server A via the data transmission / reception unit 13, the similarity calculation unit 16 obtains the data number and record number from the server A. Is written in the storage unit 19 in correspondence with the server identification information for identifying. In addition, the similarity calculation unit 16 initializes the number of checks A stored in the storage unit 19 to 0 to be used for determining whether or not the check of the number of reference data described later has been completed.

Next, the similarity calculation unit 16 reads the reference data number and the check number A from the storage unit 19, and determines whether or not the check number A is less than the reference data number (step S3).
At this time, when the number of checks A is less than the number of reference data, the similarity calculation unit 16 advances the process to step S4, and when the number of checks A is equal to or greater than the number of reference data, ends the equalization process.

Next, the similarity calculation unit 16 selects a record number to which the end flag indicating the end of processing is not given from the record number of the reference data stored in the storage unit 19, and refers to the storage unit 19 with this reference number. By writing an end flag corresponding to the record number of the data, an end flag is given to this reference data (step S4).
Then, the similarity calculation unit 16 transmits a search signal for requesting reading and transmission of reference data to the server A through the data transmitting / receiving unit 13 together with the record number of the reference data to which the end flag is added.
When the server A receives the search signal for transmitting the reference data from the similarity calculation unit 16, the server A reads the reference data of the record number given to the search signal and transmits the read reference data to the analysis server 1.

  When the reference data is supplied via the transmission / reception unit 13, the similarity calculation unit 16 deletes the previously written reference data from the storage unit 19, writes the supplied reference data to the storage unit 19, and the number of checks A Is incremented (1 is added) (step S5).

Next, the similarity calculation unit 16 accesses the server B via the data transmission / reception unit 13 to obtain the number of data in the target database provided in the server B (step S6). That is, the analysis server 1 transmits to the server B a request signal for obtaining the number of data stored in the storage database (that is, the number of records) and the record number indicating the position where each data is stored. When this request signal is received via the data transmission / reception unit 13 and the network, the server B acquires the number of records, which is the number of stored data stored in the storage database, and the record number of each data. The number of stored data and the record number of the data are transmitted to the analysis server 1.
Then, when the storage data number and the record number of the data stored in the target database are supplied from the server B via the data transmission / reception unit 13, the similarity calculation unit 16 receives the stored data number and the record number from the server B. Write to the storage unit 19 in correspondence with the server identification information for identifying B. In addition, the similarity calculation unit 16 initializes the number of checks B stored in the storage unit 19 to 0 for use in determining whether or not the check of the number of data described later has been completed.

Next, the similarity calculation unit 16 reads out the accumulated data number and the check number B from the storage unit 19, and determines whether or not the check number B is less than the accumulated data number (step S7).
At this time, the similarity calculation unit 16 advances the process to step S8 when the check number B is less than the accumulated data number, and advances the process to step S3 when the check number B is equal to or greater than the accumulated data number.

Next, the similarity calculation unit 16 selects a record number that is not given an end flag indicating the end of processing from the record numbers of the accumulated data stored in the storage unit 19, and stores this accumulation in the storage unit 19. By writing an end flag corresponding to the record number of the data, an end flag is given to the accumulated data (step S8).
Then, the similarity calculation unit 16 transmits a search signal for requesting reading and transmission of stored data to the server B through the data transmitting / receiving unit 13 together with the record number of the stored data to which the end flag is added.
When the server B receives the search signal for transmitting the stored data from the similarity calculation unit 16, the server B reads the stored data of the record number given to the search signal and transmits the read stored data to the analysis server 1.

  When the accumulated data is supplied via the transmission / reception unit 13, the similarity calculation unit 16 deletes the previously written accumulated data from the storage unit 19, writes the supplied accumulated data into the storage unit 19, and checks the number of checks B Is incremented (step S9).

  Next, the similarity calculation unit 16 reads the reference data and accumulated data stored in the storage unit 19, and calculates the similarity SCORE_ALL of the reference data and accumulated data (step S10). Details of the similarity calculation processing performed by the similarity calculation unit 16 will be described later.

  Then, the similarity calculation unit 16 determines whether the obtained similarity SCORE_ALL exceeds a preset threshold (step S11). If the similarity SCORE_ALL exceeds a preset threshold, the process proceeds to step S12. On the other hand, if the similarity SCORE_ALL is less than or equal to a preset threshold value, the process proceeds to step S7.

  Next, the rewrite item detection unit 17 reads each variable name (item_name, comment, price, keyword, Jan_code, bin_kbn, Initial) of the storage item data constituting the storage data from the storage 19 and is set in the storage unit 19. (Step S12), and the process proceeds to step S13.

  Next, the rewrite item detection unit 17 checks whether or not there is a variable name of the accumulated item data in the rewrite list of the storage unit 19 (step S13). If no variable name exists in the write list, the process is performed in step S14. On the other hand, if any variable name exists in the write list, the process proceeds to step S14.

Assuming that the process of changing the accumulated data and the reference data is completed, the data content updating unit 14 reads the accumulated data stored in the storage unit 19 together with the record number, and adds the read accumulation number and record number to the rewrite signal. Similarly, the reference data stored in the storage unit 19 is read together with the record number, and the read reference number and record number are added to the rewrite signal and transmitted to the server A (step S14).
Next, the server B rewrites the accumulated data of the record number given to the rewrite signal in the accumulation database with the accumulated data given to the rewrite signal by the rewrite signal received from the analysis server 1.
In addition, the server A rewrites the reference data of the record number given to the rewrite signal in the reference database with the reference data given to the rewrite signal by the rewrite signal received from the analysis server 1.

  Then, the rewrite item detection unit 17 reads one of the existing variable names from the rewrite list in the storage unit 19 (step S15), deletes the read variable name from the rewrite list in the storage unit 19, and the process proceeds to step S16. Proceed.

  Next, the rewrite item detection unit 17 reads the update information of the accumulation item data corresponding to the read variable name from the accumulation data in the storage unit 19 (step S16), and advances the processing to step S17.

When the update information is read, the rewrite item detection unit 17 determines whether the information described as the update information is non-target, overwrite, or union as the type of change (step S17). Here, the rewrite item detection unit 17 outputs the read type and the corresponding variable name to the data content update unit 14.
Then, the data content update unit 14 advances the process to step S13 without performing the data rewriting process when the supplied type is non-target, and advances the process to step S18 when the type is overwritten. If it is a union, the process proceeds to step S19.

FIG. 6 is a diagram illustrating a process in which the data content update unit 14 updates the content of the accumulated data in the accumulated database in FIG. 4 with the information of the reference data in the reference database.
The variable names “comment” and “price” are not subject to change because update information is not targeted.
The variable names item_name, jan_code, bin_kbn, and initial are changed by overwriting the stored item data with the reference item data because the update information is overwritten. Here, in the case of the variable name bin_kbn, since the information of the reference item data and the accumulated item data is the same before the change, it does not seem to be changed even if overwritten.
Further, since the update information of the variable name keyword is a union, both the accumulated item data and the reference item data take the OR of each other word, and the union processing is performed.

  When the update information is overwritten, the data content update unit 14 reads the reference item data corresponding to the variable name from the reference data detected as the same data in the storage unit 19, and uses the reference item data to store the same in the accumulated data. The accumulated item data of the variable name is overwritten (step S18). In addition, the data content update unit 14 advances the process to step S13.

When the update information is a union, the data content update unit 14 divides and extracts each word by the symbol # from the word string delimited by the symbol # of the stored item data, and writes the extracted word to the word list.
Subsequently, the data content update unit 14 divides and extracts each word by the symbol # from the word string delimited by the symbol # of the reference item data, writes the extracted word in the word list (step S19), and performs processing To step S20.

  Next, the data content update unit 14 detects in the storage unit 19 whether or not there are a plurality of the same words from the word list. If there are a plurality of the same words, any one of the plurality of words is selected. One other word is deleted (step S20), and the process proceeds to step S21.

  Then, the data content update unit 14 sequentially reads the words existing in the word list of the storage unit 19 one by one, and concatenates the words by inserting the symbol # between the sequentially read words (step S21). Then, the process proceeds to step S22.

Next, the data content updating unit 14 overwrites the concatenated word string on the corresponding stored item data of the stored data stored in the storage unit 19 and overwrites the corresponding reference item data of the reference data ( Step S22), the process proceeds to step S13.
In the present embodiment, in order to simplify the description, the uniformization of accumulated data has been described by the configuration with the two service servers 21 and 22. The analysis server 1 also searches for a database serving as a reference database from these databases even when there are a plurality of service servers of three or more (when configured from service servers 21 to 2n). Then, the analysis server 1 uses the storage database that has not become the reference database as a target database, and performs the above-described homogenization processing of the stored data among the databases sequentially for each database using the reference database of the reference database.

Next, the operation of the reference database selection process performed by the reference database setting unit 15 in the analysis server 1 of this embodiment will be described with reference to FIG. FIG. 8 shows that the reference database setting unit 15 determines which one of a plurality of databases is the most reliable reference database based on each reliability standard of the standard table (shown in FIG. 7) stored in the reliability data storage unit 18. FIG.
The service servers managed by the analysis server 1 are assumed to be two service servers 21 and 22, and the reference database selection processing by the reference database setting unit 15 will be described below. At this time, it is assumed that the service server 21 is the server A and the service server 22 is the server B in the previous determination.

  The reference database setting unit 15 reads the variable names Validity, Organization, Specialty, Person, Access, Update, and Advertisement of each criterion from the criterion table of the reliability data storage unit 18 and sequentially writes them to the selection table of the storage unit 19 (steps). S101), the process proceeds to step S2. At this time, the reference database setting unit 15 initializes an evaluation value VALUE_ALL indicating which database has high reliability to 0.

  Next, the reference database setting unit 15 determines whether or not any variable name exists in the selection table of the storage unit 19 (step S102). If no variable name remains, the process proceeds to step S102. On the other hand, if any variable name exists, the process proceeds to step S105.

  Then, the reference database setting unit 15 determines whether the evaluation value VALUE_ALL is positive or negative (whether it is 0 or more) (step S103). When the evaluation value VALUE_ALL is positive, the database DB1 provided in the service server 21 that is the server A is set to be more reliable than the database DB2 provided in the service server 22 that is the server B, and the database DB1 is set as the reference database. Then, the reference database setting process is terminated without changing the state in which the database DB2 is set as the target database. On the other hand, when the evaluation value VALUE_ALL is negative, it is assumed that the database DB1 provided in the service server 21 that is the server A is lower in reliability than the database 22 provided in the service server 22 that is the server B. And the database DB1 is set as the target database (switching between the server A and the server B), and the reference database setting process is terminated.

  If any variable name exists in the selection list, the reference database setting unit 15 reads any variable name existing in the selection list in the storage unit 19 (step S105), and advances the process to step S106. At this time, the reference database setting unit 15 deletes the read variable name from the selection list in the storage unit 19.

When the variable name is read, the reference database setting unit 15 reads the information collected by the reliability standard corresponding to the variable name from each of the reliability tables corresponding to the databases DB1 and DB2 of the reliability data storage unit 18 ( Step S106).
Here, the reference database setting unit 15 determines whether the information read and collected from the reliability table is a numerical value or presence / absence information indicating whether or not the reliability standard is satisfied based on the data type ( In step S107), if the data type is a numerical value, the process proceeds to step S108. On the other hand, if the data type is T / F, the process proceeds to step S114. In FIG. 7, as numerical values, there are legitimacy (the number of links of external links), the number of accesses, the number of updates, and the advertisement area (the ratio indicating the ratio of the advertisement display area to the total display area of the site).

  When the data type is a numerical value, the reference database setting unit 15 assigns the numerical value of the information collected in the database DB1 to the variable Dat_A, and similarly substitutes the numerical value of the information collected in the database DB2 into the variable Dat_B (step S108). The process proceeds to step S109.

  Next, the reference database setting unit 15 determines whether or not the numerical value assigned to the variable Dat_A is equal to the numerical value assigned to the variable Dat_B (step S109), and the numerical value assigned to the variable Dat_A and the variable If the numerical value assigned to Dat_B is equal, the process proceeds to step S111. On the other hand, if the numerical value assigned to variable Dat_A is different from the numerical value assigned to variable Dat_B, the process proceeds to step S110.

  When the numerical value assigned to the variable Dat_A is equal to the numerical value assigned to the variable Dat_B, the reference database setting unit 15 assigns 0 to the variable VALUE (step S111), and the process proceeds to step S121.

  When the numerical value assigned to the variable Dat_A is different from the numerical value assigned to the variable Dat_B, the reference database setting unit 15 determines whether or not the numerical value assigned to the variable Dat_A is larger than the numerical value assigned to the variable Dat_B. If the numerical value assigned to the variable Dat_A is larger than the numerical value assigned to the variable Dat_B, the process proceeds to step S112. On the other hand, if the numerical value assigned to the variable Dat_A is smaller than the numerical value assigned to the variable Dat_B, the process proceeds to step S113.

  When the numerical value assigned to the variable Dat_A is larger than the numerical value assigned to the variable Dat_B, the reference database setting unit 15 assigns 1 to the variable VALUE (step S112), and the process proceeds to step S121.

  If the numerical value assigned to the variable Dat_A is smaller than the numerical value assigned to the variable Dat_B, the reference database setting unit 15 assigns −1 to the variable VALUE (step S113), and the process proceeds to step S121.

  When the data type is T / F in step S107, the reference database setting unit 15 substitutes the letter T or F indicating presence / absence as information collected in the database DB1 into the variable Dat_A, and similarly collects information in the database DB2. Then, the numerical value of T or F indicating presence / absence is substituted into the variable Dat_B (step S114), and the process proceeds to step S115. In FIG. 7, T / F includes organization information, expertise, and related persons as reference names.

  Next, the reference database setting unit 15 determines whether or not the character assigned to the variable Dat_A is equal to the character assigned to the variable Dat_B (step S115), and the character and variable assigned to the variable Dat_A are determined. If the character assigned to Dat_B is equal, the process proceeds to step S116. If the character assigned to variable Dat_A and the character assigned to variable Dat_B are different, the process proceeds to step S117.

  When the character assigned to the variable Dat_A is equal to the character assigned to the variable Dat_B, the reference database setting unit 15 determines whether or not the character written to the variable Dat_B is T (True) (step) S116) If the character written in the variable Dat_B is not T, the process proceeds to step S118. On the other hand, if the character written in the variable Dat_B is T, the process proceeds to step S119.

  When the character assigned to the variable Dat_A is not equal to the character assigned to the variable Dat_B, the reference database setting unit 15 determines whether or not the character written to the variable Dat_B is T (True) ( In step S117), if the character written in the variable Dat_B is T, the process proceeds to step S120. On the other hand, if the character written in the variable Dat_B is not T, the process proceeds to step S119.

  If the character written in the variable Dat_B in step S116 is not T, the reference database setting unit 15 assigns 1 to the variable VALUE (step S118), and the process proceeds to step S121.

  If the character written in the variable Dat_B in step S116 is T, or if the character written in the variable Dat_B in step S117 is not T, the reference database setting unit 15 substitutes 0 for the variable VALUE (step S119). Then, the process proceeds to step S121.

  Next, the reference database setting unit 15 reads the weighting coefficient assigned to the reliability standard corresponding to the variable name from the standard table of the reliability data storage unit 18, and multiplies the variable VALUE by the read weighting coefficient. The result is added to the evaluation value VALUE_ALL, each reliability criterion in the criterion table is integrated (step S121), and the process proceeds to step S102.

  If the character written in the variable Dat_B in step S117 is T, the reference database setting unit 15 assigns −1 to the variable VALUE (step S120), and the process proceeds to step S121.

In the present embodiment, the databases provided in the service servers are handled in parallel, that is, which database is highly trusted by the user when performing the process of uniformizing the accumulated data stored in the databases. And the stored data stored in the other target database is changed according to the reference data stored in the reference database, using the most reliable database as a reference database.
As another configuration, a reference database to be referred to in the analysis server 1 may be set in advance in a plurality of database servers under management, and stored data stored in the reference database may be used as reference data.

Next, the operation of the calculation processing of the similarity between the reference data and the accumulated data performed by the similarity calculation unit 16 in the analysis server 1 of the present embodiment will be described with reference to FIGS. In FIG. 9, the similarity calculation unit 16 calculates the similarity between the reference data (server A reference database) and the storage data (server B target database) in FIG. It is a flowchart which shows the operation example which calculates the similarity for every item and determines whether it is the same data from the sum total of the similarity for every item. FIG. 10 is a diagram illustrating a data configuration of the reference data and the accumulated data for explaining the calculation process I of the similarity between the reference data of the reference database and the accumulated data of the target database. FIG. 11 is a diagram for explaining the calculation process II of the similarity for each item of data. The calculation formula for the similarity for each variable name, the weighting number by which each similarity is multiplied, and the SCORE calculated as an example. The numerical value of is described.
The service servers managed by the analysis server 1 are assumed to be two service servers 21 and 22, and the reference database selection processing by the reference database setting unit 15 will be described below. At this time, it is assumed that the service server 21 is the server A and the service server 22 is the server B in the previous determination.

  The similarity calculation unit 16 reads out the variable names item_name, comment, price, Keyword, jan_code, bin_kbn, and Initial of all items of the accumulation item data forming the accumulation data stored in the storage unit 19 (Step S201). Write to the calculation table set in the storage unit 19.

  Next, the similarity calculation unit 16 determines whether any variable name exists in the calculation table of the storage unit 19 (step S202), and any variable name exists in the calculation table. If there is no variable name in the calculation table, the calculation of the similarity is terminated, and the similarity SCORE_ALL at this time is output as the similarity between the reference data and the accumulated data. To do.

  If any variable name exists in the calculation table, the similarity calculation unit 16 reads the variable name of any item from the calculation table in the storage unit 19 (step S203), and advances the process to step S204. At this time, the similarity calculation unit 16 deletes the variable name read from the calculation table of the storage unit 19.

Next, the similarity calculation unit 16 determines whether or not the accumulated item data of the item corresponding to the read variable name exists (step S204), and if the accumulated item data does not exist, the process proceeds to step S204. The process proceeds to S205, and if the accumulated item data exists, the process proceeds to step S206.
For example, as shown in FIG. 10, since there is no variable name jan_code and Initial accumulated item data in the accumulated data, in the case of this variable name, the similarity calculation unit 16 proceeds with the process to step S205. Become. For this reason, as shown in FIG. 11, the similarity calculation unit 16 substitutes 1 for the variable SCORE in step S205.
On the other hand, since the accumulated item data exists in the other variable names item_name, comment, price, Keyword, and bin_kbn, the similarity calculation unit 16 proceeds with the process to step S206.

  When the accumulation item data does not exist, the similarity calculation unit 16 substitutes 1 for the variable SCORE (step S205), and the process proceeds to step S212.

  When the accumulated item data exists, the similarity calculation unit 16 reads the meta information of the variable name item from the accumulated data in the storage unit 19 (step S206), and advances the process to step S207.

  Next, the similarity calculation unit 16 determines whether the read meta information indicates a numerical value, an ID, a single word, a character string, or a multiple word (step S207). If the meta information is a numerical value, If the meta information is an ID, the process proceeds to step S213. If the meta information is a single word, the process proceeds to step S217. If the meta information is a character string, the process proceeds to step S222. If the meta information is a multiple word, the process proceeds to step S224.

  When the read meta information is a numerical value, the similarity calculation unit 16 substitutes the numerical value of the reference item data of the reference data of the database DB1 for the variable Dat_A, and similarly, the numerical value of the stored item data of the stored data of the database DB2 is the variable Dat_B. (Step S208), and the process proceeds to step S209.

  Next, the similarity calculation unit 16 determines whether or not the numerical value assigned to the variable Dat_A exceeds the numerical value assigned to the variable Dat_B (step S209), and the numerical value assigned to the variable Dat_A is the variable. If it is smaller than the value assigned to Dat_B, the process proceeds to step S210. On the other hand, if the value assigned to variable Dat_A exceeds the value assigned to variable Dat_B, the process proceeds to step S211.

  When the numerical value assigned to the variable Dat_A is smaller than the numerical value assigned to the variable Dat_B, the similarity calculation unit 16 assigns the result of dividing the variable Dat_A by the variable Dat_B to the variable SCORE (step S210). The process proceeds to step S212.

When the numerical value assigned to the variable Dat_A exceeds the numerical value assigned to the variable Dat_B, the similarity calculation unit 16 assigns the result of dividing the variable Dat_B by the variable Dat_A to the variable SCORE (step S211). The process proceeds to step S212.
For example, in the case of the variable name price, the numerical value of the reference item data is 350, and the numerical value of the accumulated item data is 348.
Therefore, the similarity calculation unit 16 performs a calculation of 348/350 as shown in FIG. 11 and substitutes the obtained 0.999... Into the variable SCORE.

  Next, the similarity calculation unit 16 reads the weighting coefficient corresponding to the variable name from the accumulated data in the storage unit 19, multiplies the read weighting coefficient by the variable SCORE, and uses the obtained multiplication result to the similarity SCORE_ALL. Then, addition is performed (step S212), the calculation is performed for each data item, the weighted variable SCORE is integrated, and the similarity is calculated.

  When the meta information is ID in step S207, the similarity calculation unit 16 substitutes ID identification information of the reference item data of the reference data of the database DB1 into the variable Dat_A, and similarly stores the accumulated item data of the accumulated data of the database DB2. The ID identification information is substituted into the variable Dat_B (step S213), and the process proceeds to step S214.

  Next, the similarity calculation unit 16 determines whether or not the variable Dat_A and the variable Dat_B are equal (step S214). If the variable Dat_A and the variable Dat_B are equal, the process proceeds to step S215. If Dat_A and variable Dat_B are not equal, the process proceeds to step S216.

  When the variable Dat_A and the variable Dat_B are equal, the similarity calculation unit 16 substitutes 1 for the variable SCORE (step S215), and the process proceeds to step S212.

  When the variable Dat_A and the variable Dat_B are not equal, the similarity calculation unit 16 substitutes 0 for the variable SCORE (step S216), and the process proceeds to step S212.

  When the meta information is a single word in step S207, the similarity calculation unit 16 reads the word data of the reference item data in the reference data and the word data of the accumulation item data in the accumulation data from the storage unit 19 (step S217). ), And the process proceeds to step S218.

  Next, the similarity calculation unit 16 determines whether or not the word of the reference item data and the word of the accumulated item data completely match (that is, whether or not they are the same) (step S219). If the word of the reference item data and the word of the accumulated item data are completely matched, the process proceeds to step S22. On the other hand, if the word of the accumulated item data is not completely matched, the process proceeds to step S221. Proceed to

  If the word of the reference item data and the word of the stored item data are completely the same, the similarity calculation unit 16 substitutes 1 for the variable SCORE (step S220), and the process proceeds to step S212.

When the word of the reference item data and the word of the stored item data do not completely match, the similarity calculation unit 16 substitutes 0 for the variable SCORE (step S221), and the process proceeds to step S212.
For example, in the case of the variable name bin_kbn, since the word “frozen” in the reference item data and the word “frozen” in the accumulated item data completely match, the similarity calculation unit 16 sets the variable SCORE in step S220. For this, 1 is substituted.

  If the meta information is a character string in step S207, the similarity calculating unit 16 stores the word data of the character string of the reference item data in the reference data and the character string data of the accumulated item data in the accumulated data from the storage unit 19. Reading (step S222), the process proceeds to step S223.

  Next, the similarity calculation unit 16 performs a morphological analysis on the character string of the read reference item data, divides it into words, writes the divided words to the similar word A list set in the storage unit 19, and similarly reads them. The character string of the stored item data is morphologically analyzed, divided into words, the divided words are written in the similar word B list set in the storage unit 19 (step S223), and the process proceeds to step S226. At this time, the similarity calculation unit 16 initializes the variable COUNT in the storage unit 19 to 0, and calculates the number of words existing in the similar word A list of the storage unit 19 and the number of words existing in the similar word B list. The identification information, which is either reference data or accumulated data, is given, and is written and stored in the storage unit 19 together with this identification information.

  Then, the similarity calculation unit 16 determines whether or not a word exists in the similar word A list (step S226), and if a word exists in the similar word A list, the process proceeds to step S227. On the other hand, if a word does not exist in the similar word A list, the process proceeds to step S231.

  When a word exists in the similar word A list, the similarity calculation unit 16 reads one of the existing words (step S227), and advances the process to step S228. At this time, the similarity calculation unit 16 deletes the word read from the similarity word A list from the similarity word A list.

  Next, the similarity calculation unit 16 compares the word read from the similarity word A list with the word existing in the similarity word B list (step S228), and advances the process to step S229.

  In the comparison result, the similarity calculation unit 16 determines whether the word read from the similarity word A list and the word existing in the similarity word B list are a complete match or a partial match with the same word part. It is determined whether there is no matching part (step SS229), and if the word read from the similarity word A list and the word existing in the similarity word B list do not completely match, processing is performed The process proceeds to step S226, and if the word read from the similarity word A list and the word existing in the similarity word B list match completely or partially, the process proceeds to step S230.

  When the word read from the similarity word A list and the word existing in the similarity word B list completely or partially match, the similarity calculation unit 16 increments the variable COUNT (step S230), and the process proceeds to step S226. Proceed to

If no word remains in the similarity word A list in step S226, the similarity calculation unit 16 stores the number of words existing in the similarity word A list and the number of words existing in the similarity word B list in the storage unit 19. A numerical value COA obtained by dividing the calculated variable COUNT by the number of words existing in the similarity word A list and a numerical value COB obtained by dividing the calculated variable COUNT by the number of words existing in the similarity word A list are calculated. To do.
Then, the similarity calculation unit 16 outputs the larger one of the calculated numerical value COA and numerical value COB as the value of the variable SCORE (step S231).

For example, if the meta information is a string variable name comment, the reference item data is <soft even if it is cooled, and it is because it is juicy and delicious. It is a long-selling character that has been loved for its unaltered taste>, and the similarity calculation unit 16 uses a morphological analysis to make it “soft”, “juicy” and “taste”. "Karaage". It is a “long seller” that is “loved” and “continues” by “taste” that does not change, and breaks it down into 10 words.
Similarly, the similarity calculation unit 16 is the accumulated item data because it is a very popular cold and soft and juicy. Morphological analysis is performed on the string of Japanese chicken bouillon and mellow umami, which makes it “soft”, “juicy” and “fried” even if it is “cool”. “Domestic” “Chicken” “bouillon” and “mellow” “umami” are broken into 11 words as “feature”>.

  Then, as shown in FIG. 11, the similarity calculation unit 16 includes four words that are completely matched or partially matched, 10 words obtained by morphological analysis of the reference item data, and morphological analysis of the accumulated item data. The number of words obtained in this way is calculated as 11, respectively. The value (0.4) obtained by dividing the number of completely matched or partially matched words by the number of words obtained by morphological analysis of the reference item data is divided by the number of words obtained by morphological analysis of the stored item data. Since it is larger than the value (0.36), 0.4 is substituted into the variable SCORE.

  In step S207, when the meta information includes multiple words, the similarity calculation unit 16 reads from the storage unit 19 the word string data of the reference item data in the reference data and the word string data of the accumulation item data in the accumulation data ( Step S224), the process proceeds to step S225.

  Next, the similarity calculation unit 16 extracts the word string of the read reference item data by using a symbol # sandwiched between words, divides the word string, and the divided words are set to the similar word set in the storage unit 19. Similarly, the word string of the stored item data read is extracted by the symbol # inserted between words, divided into words, and the divided word is set to the similar word B set in the storage unit 19. Writing to the list (step S225), the process proceeds to step S226. At this time, the similarity calculation unit 16 initializes the variable COUNT in the storage unit 19 to 0, and calculates the number of words existing in the similar word A list of the storage unit 19 and the number of words existing in the similar word B list. The identification information, which is either reference data or accumulated data, is given, and is written and stored in the storage unit 19 together with this identification information.

For example, when the meta information is a variable name item_name having a plurality of words, the similarity calculation unit 16 uses “#Company A” by the symbol # from the reference item data <Company A>#Bird raising # 6 #Frozen> , Extract four words “bird fried”, “with 6” and “frozen”.
Similarly, the similarity degree calculation unit 16 has three items of “Company A”, “Tori-no-Koriage” and “6” from the symbol #, based on the accumulated item data of the accumulated data <Company A #Takera # 6>. Extract words.

Then, the similarity calculation unit 16 detects “Company A” as a complete match, and detects “bird raising” and “torino raising”, “six pieces”, and “6 pieces” as partial matches. As a result, the variable COUNT becomes 3.
For this reason, since the number of words extracted from the word string of the reference item data is four and the number of words extracted from the word string of the accumulated item data is three, the similarity calculation unit 16 has a complete match or partial A value (0.75) obtained by dividing the number of matched words by the number of words extracted from the word string of the stored item data (1.0) and the number of words extracted from the word string of the reference item data (0.75) ) Is larger, 1 is substituted into the variable SCORE.

  With the above-described configuration, the analysis server 1 according to the present embodiment sets, as a reference database, a database that is determined to be most reliable in search among a plurality of different databases under management (DB1 to DBn in FIG. 2). Then, other databases are set as target databases to be updated. The similarity between the data stored in each of the reference database and the target database is obtained, and the data whose similarity is higher than the set threshold is determined as the same data. By updating the data in the target database by referring to the data in the reference database, the contents of the data between the databases can be made uniform, and it is possible to easily maintain a database with high reusability of the accumulated data. It becomes.

  Further, a program for realizing the functions of the data content update unit 14, the reference database setting unit 15, the similarity calculation unit 16, and the rewrite item detection unit in FIG. 3 is recorded on a computer-readable recording medium. The program recorded in the above may be read into a computer system and executed to perform a uniform process between a plurality of data. Here, the “computer system” includes an OS and hardware such as peripheral devices.

Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design and the like within a scope not departing from the gist of the present invention.

  The embodiment of the present invention can be used in the field of information retrieval for a plurality of databases.

DESCRIPTION OF SYMBOLS 1,100 ... Analysis server 21, 22, 2n, 102, 105 ... Service server DB1, DB2, DBn ... Database 11 ... Control part 12 ... Data input part 13 ... Data transmission / reception part 14 ... Data content update part 15 ... Reference database setting Unit 16: Similarity calculation unit 17 ... Rewrite item detection unit 18 ... Reliability data storage unit 19 ... Storage unit 20 ... Display device

Claims (12)

The reference data stored in the reference database for reference is compared with the stored data stored in the storage database for storage, and the similarity between the stored data and the reference data is obtained. A similarity calculator that detects the accumulated data in which the similarity exceeds a preset threshold as similar data;
A data content update unit for updating the stored data detected as the same data with the data content of the stored data. Each of the accumulated data and the reference data is composed of a plurality of configuration data, and whether or not to update the information for each configuration data is set in advance,
The database according to claim 1, further comprising a rewrite item detecting unit that determines, for each of the configuration data, whether or not to update the accumulated data based on the data content of the reference data based on the availability information. Control device.   When the availability information set in the configuration data is acceptable, the configuration data of the stored data that is the same data as the reference data is changed by the configuration data corresponding to the reference data, The configuration data of the accumulated data that is the same data as the reference data is not changed by the configuration data corresponding to the reference data when the availability information is NO. 3. The database control device according to 2.   If the availability information is acceptable, the configuration data is overwritten with the corresponding configuration data in the data of the reference database, or the configuration data of the data and the corresponding configuration of the data in the reference database 4. The database control apparatus according to claim 3, wherein a change type as to whether it is a union with data is set.   The similarity calculation unit is configured to determine the similarity between the configuration data of the reference data and the configuration data of the stored data, which is set for each data type of the configuration data. 5. The degree of similarity of each of the configuration data in the data with the storage database is calculated, and the sum of the calculated degrees of similarity is used as the degree of similarity of the data. The database control device according to one item. Meta information included in the configuration data indicates a data type,
The similarity calculation unit has a calculation method table indicating a relationship between the related data type and a similarity calculation method, extracts meta information from the configuration data, and calculates a similarity corresponding to the extracted meta information The database control apparatus according to claim 5, wherein a method is selected from the calculation method table, and the similarity for each of the configuration data is calculated.   The similarity calculation unit is configured to multiply the similarity obtained by the similarity calculation method by a weighting coefficient set in the configuration data to obtain the similarity of the configuration data. Item 7. The database control device according to Item 6. A reliability data storage unit that stores, for each database, reliability data indicating the reliability of the database composed of a plurality of items updated every certain period;
A reference database setting unit that obtains the reliability of each database based on the reliability data and sets the database with the highest reliability in the reference database. The database control device according to any one of the above.   The reliability data is any one of at least the number of external links to the server managing the database, the presence or absence of access information, the number of accesses by the user, the number of site updates, the area of the advertisement display displayed on the site, and the ratio Alternatively, the database control apparatus according to claim 8, wherein the database control apparatus is a plurality of combinations. A database control method for operating a database control device that uniformizes the data content of stored data stored in each database among a plurality of databases,
The similarity calculation unit compares the data contents of the reference data stored in the reference database for reference and the stored data stored in the storage database for storage, and the similarity between the stored data and the reference data A similarity calculation step of detecting the accumulated data in which the obtained similarity exceeds a preset threshold as similar data;
A data content update process in which a data content update unit updates the accumulated data detected as the data with the data content of the accumulated data. A program that causes a computer to execute the function of a database control device that uniformizes the data content of stored data stored in each database among a plurality of databases.
The similarity calculation unit compares the data contents of the reference data stored in the reference database for reference and the stored data stored in the storage database for storage, and the similarity between the stored data and the reference data A degree of similarity calculation process for detecting the accumulated data in which the obtained degree of similarity exceeds a preset threshold as similar data;
A computer-executable program, comprising: a data content update process in which a data content update unit updates the stored data detected as the same data with the data content of the stored data. A recording medium that records a program that causes a computer to execute the function of a database control device that equalizes the data content of accumulated data accumulated in each database among a plurality of databases.
The similarity calculation unit compares the data contents of the reference data stored in the reference database for reference and the stored data stored in the storage database for storage, and the similarity between the stored data and the reference data A degree of similarity calculation process for detecting the accumulated data in which the obtained degree of similarity exceeds a preset threshold as similar data;
A recording medium recording a computer-executable program having a data content update process in which a data content update unit updates the stored data detected as the same data with the data content of the stored data.

Images