JP6949449B2 - Data search system and data search program - Google Patents

Description

The present invention relates to a data search system and a data search program.

Conventionally, there has been a problem that an extremely large amount of time is required when performing an ambiguous search in a database system such as a relational database having a large amount of data.

Patent Document 1 discloses a document search device capable of shortening the search time by using the state information of high-speed search preprocessing using a full-text search engine and an RDB (relational database).

Specifically, a pattern matching unit and a full-text search unit are provided for a plurality of documents and folders, the type and state of the search target are determined, and either the pattern matching unit or the full-text search unit is used. It is a search. The invention of Patent Document 1 is characterized in that the pattern matching unit can be searched immediately but the search time is long, and the full-text search unit is characterized in that the registration takes time but the search time is high. In view of this, these are switched and each is used in an advantageous situation.

In Patent Document 2, a full-text search engine capable of communicating with a plurality of clients adds a control character (for example, a delimiter) to the identification code (user ID) of the client to which the full-text search engine has the right to access the document at the time of document registration. The character string is stored together with the above document, and the full-text search engine adds the character string in which the above control character is added to the identification code of the client who requested the access to the search term when the document is requested to access the document and performs a full-text search. Is disclosed.

Further, in the invention of Patent Document 2, the full-text search engine has a column for storing the attribute value of the document separately from the document body, and the full-text search engine uses the identification code and the control character of the client at the time of document registration. A character string consisting of is stored in a column as an attribute value of a document, and a character string in which a control character is added to the identification code of the client requested to access the document is searched in full text as a search term for the column. As a result, access control can be performed at the same time as full-text search.

Further, Patent Document 3 discloses a search engine including a crawler that creates a document index for full-text search by crawling a document DB holding document data describing each issue at a predetermined timing as a financial information retrieval system. Has been done. This search engine further has a search processing unit that responds as a search result to a predetermined number of high-ranking records related to document data matching from the document index in response to a search request received from a sales terminal. It has a DB search unit that directly performs a search process on a document DB instead of a search by a search engine when a keyword is not specified in the search request received from the terminal.

According to the invention of Patent Document 3, even when a keyword is not specified at the time of a search by a search engine, data that can be included in a predetermined number of high-ranking items under the sorting conditions for the search result is displayed without omission. It will have the effect of becoming.

Further, Patent Document 4 includes a data type analysis unit that analyzes data structure features (data types, etc.) of a user-specified search keyword input by the input unit, and a search target stored in a relational database. A search target column determination including a search target column detection unit that detects a column that matches the data structure characteristics (data type, etc.) of the analyzed search keyword as the search target column among the columns of the table. The device is disclosed.

The invention of Patent Document 4 is to improve the response performance at the time of full-text search by dynamically narrowing down the columns to be the target of full-text search based on the data structure characteristics of the search keyword.

Japanese Unexamined Patent Publication No. 2006-79423 JP-A-2009-169736 Japanese Unexamined Patent Publication No. 2015-185013 JP-A-2010-67213

An object of the present invention is to provide a data search system and a data search program that enable a faster search than the above-mentioned search system.

The data search system according to the present invention searches for a database in which a plurality of tables of managed data in one unit are accumulated and all of the plurality of tables in the database with search information, and is one unit of the managed data. A first index table in which all the tables are specified, unique values for all the specified tables are used as the primary key information, and the search target information in the content data of the corresponding table is assigned to this primary key information as attribute information. Is given, and a first crawler collecting means for generating a first index file that summarizes the first index table and a keyword to be searched are given, the first index file is searched and the relevant first index file is searched. The primary key information acquisition means for detecting the first index table having data corresponding to the keyword and obtaining the primary key information of the first index table, and the primary key information obtained by the primary key information acquisition means. A database search means for searching a database and extracting data corresponding to the keyword from the obtained corresponding table, a display means for displaying information, and a display for displaying based on the data extracted by the database search means on the display means. It is characterized by having a control means.

The block diagram which shows the structure of embodiment of the data search system which concerns on this invention. The figure which shows an example of the contents of the database used in embodiment of the data search system which concerns on this invention. FIG. 5 is a diagram showing an example of a process of creating a first index table in a first index file from a database table in an embodiment of a data retrieval system according to the present invention. The figure which shows an example of the contents of the attached file stored in the file apparatus used in embodiment of the data search system which concerns on this invention. FIG. 5 is a diagram showing an example of a process of creating a second index table in a second index file from an attached file of a file device in an embodiment of a data retrieval system according to the present invention. The flowchart which shows the operation of the embodiment of the data search system which concerns on this invention.

Hereinafter, embodiments of the data search system and the data search program according to the present invention will be described with reference to the accompanying drawings. In each figure, the same components are designated by the same reference numerals, and duplicate description will be omitted. FIG. 1 shows a configuration diagram of an embodiment of a data retrieval system according to the present invention. The data search system according to the embodiment includes a database 300 in which various data are stored, and a full-text search engine 500 that searches all sentences in the database 300. As the database 300, for example, a relational database can be adopted.

Here, the managed data accumulated in the database 300 is the data of the product catalog, and for example, as shown in FIG. 2, the tables D11, D12, D13, ..., D1n of one unit of managed data are It has been accumulated. One table has a structure in which a plurality of necessary items are arranged for unique values. In the present embodiment, the unique value is the product number arranged at the head of the table, and in each table D11, D12, D13, ..., D1n, as item data, "product name" and "product" are displayed. It is lined up with "name kana", "packaging", "handling start date", "handling end date", and so on. The order of the data in the items is only an example.

The full-text search engine 500 includes a first crawler collecting means 503. The first crawler collecting means 503 searches the database 300, identifies a table that is one unit of the managed data, and uses unique values for all the identified tables as primary key information, and this primary key. A first index table is created in which the search target information in the content data of the corresponding table is assigned to the information as attribute information, and a first index file 501 that summarizes the first index table is generated.

As already explained, in the table D11 of the database 300, the item number and the item data are "product name", "product name kana", "packing style", "handling start date", and "handling end date". Since ... -From the data, the information of the item (column) of the character string corresponding to the search target information specified in SQL (Structured Query Language) is searched and assigned as the attribution information to create the first index table.

FIG. 3 shows a process of creating the first index table D41 from the table D11. Here, it is shown that the "product number", "product name", and "product kana" correspond to the search target information specified by SQL. The same search is performed for the tables D12 to D1n, and an index table is created corresponding to the table in which the information corresponding to the search target information specified by SQL is included in the item (column). Therefore, index tables are not created for all of the tables D11 to D1n. For example, since the table D12 does not include the "product number", "product name", and "product kana" of the search target information specified by SQL, the index table corresponding to this table is not created. The several index tables created as described above are all combined into one to be the first index file 501.

In the present embodiment, the file device 400 in which the attached files under the data directory of the database 300 are stored is provided. For example, one attached file D21 in the file device 400 is as shown in FIG. 4, and is an attached file under the directory of the table D11 shown in FIG. Then, the unique value of the attached file D21 is the same "product number" as the table D11 to indicate the directory subordinate of the table D11, and a unique "serial number" is added to the attached file D21. There is. In the attached file D21, in addition to the above-mentioned unique value "product number_serial number", data such as a pamphlet and an instruction manual associated with this product number are stored. Similar to the attached file D21, a plurality of attached files are accumulated in the file device 400, and the attached file has a unique "product number_serial number", as well as brochures and handling associated with this product number. Data such as instructions are saved. These attached files are stored under the directories of the tables D11, D12, D13, ..., D1n in which the database 300 is stored, respectively. The attached file is not necessarily associated with all of the tables D11, D12, D13, ..., D1n, and there is a table to which the attached file is not associated. Further, as is clear from the fact that the "product number_serial number", which is the unique identification information that is the identifier of the attached file, is the "serial number", there are a plurality of items for one table in the database 300. The attached file may exist. In this case, the "serial number" part is "01", "02", ....

The full-text search engine 500 is provided with a second crawler collecting means 504. The second crawler collecting means 504 searches the file device 400 and creates a second index table in which the required data of the attached file is assigned to the identification information with a unique value, and the second index table is created. It generates a second index file 502 that summarizes the index table.

As described above, the file device 400 stores attached files D21, etc., and the attached file has a unique "product number_serial number" and a pamphlet associated with this product number. And data such as instruction manuals are saved. The second crawler collecting means 504 is a character string corresponding to the search target information specified in SQL (Structured Query Language) from data such as a brochure or an instruction manual associated with "product number_serial number". A second index table is created by assigning the data of. The search target information specified in this SQL (Structured Query Language) may be different from the search target information used when creating the first index table. The first crawler collecting means 503 and the second crawler collecting means 504 can be configured to perform a search by either morphological analysis or N-Gram.

FIG. 5 shows a process of creating a second index table D42 from the attached file D21 of the file device 400. Here, the search target information specified by SQL indicates that data such as a pamphlet or an instruction manual is applicable. The second index table D42 uses "item number_serial number", which is a unique value of the attached file D21, as identification information, and the pamphlet or the pamphlet corresponding to the search target information specified by the SQL of the attached file corresponding to this identification information. Create a second index table to which data such as instruction manuals are assigned. The same search is performed for the attached file (not shown) other than the attached file D21 in the file device 400, and when the data corresponding to the search target information specified by the SQL hits, it is the identification information of this attached file. An index table is created corresponding to "Product number_Sequential number". If the data corresponding to the search target information specified by SQL is not hit, the index table is not created. Therefore, the index table is not created for all the attached files in the file device 400. The several index tables created as described above are all combined into one to form a second index file 502.

In this embodiment, the search device 200 is provided. The meaning of "book" in the search device 200 is that the search is performed by the full-text search engine 500 as well, so that the search after the original search request arrives is performed in the search device 200. Is. A search request and a keyword are given to the search device 200 from the search terminal 101. As the search terminal 101, a personal computer or workstation connected by a network or the like, a mobile mobile terminal such as a mobile phone or a smartphone, or the like can be adopted.

The search device 200 is provided with a primary key information acquisition means 205 and a database search means 201. When a keyword to be searched is given, the primary key information acquisition means 205 searches the first index file 501, detects a first index table having data corresponding to the keyword, and detects the first index table. It asks for the primary key information of the index table. Specifically, the primary key information acquisition unit 205 obtains the primary key information by giving a search request to the full-text search engine 500 to search the first index file 501. Since a plurality of first index tables to which the attribute information belongs are stored in the first index file 501 with the "product number" as the primary key information, a search is performed using keywords for all of the first index tables. Then, the first index table including the corresponding keyword is obtained, and the "product number" which is the primary key information is obtained. Therefore, when the search of the first index file 501 by the full-text search engine 500 is completed, some primary key information "product number" is obtained, or there is no attribute information that hits the keyword, so that "product number" is obtained. Is not being sought. This information is sent to the primary key information acquisition means 205.

The database search means 201 searches the database 300 based on the primary key information obtained by the primary key information acquisition means 205, and extracts data corresponding to the keyword from the obtained corresponding table. That is, when the database search means 201 searches the database 300 based on the primary key information, it quickly and surely arrives at the table of the database 300 in which the corresponding data exists. Therefore, it is desired to correspond to the keyword from this table. Data can be retrieved.

The search device 200 is provided with display control means 206. The data obtained by the database search means 201 is sent to the display control means 206. The display control means 206 sends a display based on the data extracted by the database search means 201 to the search terminal 101 so that the display means performs the display.

The display control means 206 is provided with a display data processing means 203 and a display processing means 204. The display data processing means 203 processes the data (data hit corresponding to the keyword) obtained by the database search means 201 as data to be displayed in a list on the search terminal 101. The display processing means 204 sends the processed data as display data that can be displayed on the display means (display such as LED) of the search terminal 101.

Further, the search device 200 is provided with identification information acquisition means 207 and attachment file search means 202. When a keyword to be searched is given, the identification information acquisition means 207 searches the second index file 502, finds a second index table having data corresponding to the keyword, and detects the second index. It asks for the identification information of the table. Specifically, the identification information acquisition unit 207 obtains the identification information by giving a search request to the full-text search engine 500 and causing the second index file 502 to be searched. Since the second index file 502 stores a plurality of second index tables in which "product number_serial number" is used as identification information and data such as brochures and instruction manuals are assigned, this second index file 502 is stored. A search is performed on all index tables using keywords, a second index table including the corresponding keyword is obtained, and the identification information "product number_serial number" is obtained. Therefore, when the search of the second index file 502 by the full-text search engine 500 is completed, some identification information "product number_serial number" is obtained, or a second index table having data that hits the keyword is obtained. Because there was no such thing, "Product number_Sequential number" was not obtained. This information is sent to the identification information acquisition means 207.

The attached file search means 202 searches the file device 400 based on the identification information obtained by the identification information acquisition means 207, and extracts data corresponding to the keyword from the obtained attached file. That is, when the file device 400 is searched based on the identification information obtained by the attached file search means 202, the table of the file device 400 in which the corresponding data exists can be reliably reached at high speed. The corresponding desired data can be retrieved.

The data extracted in this way is sent to the display control means 206, and the display control means 206 sends a display based on the data extracted by the database search means 201 to the search terminal 101, and the display means displays the display. Try to do it.

The display data processing means 203 processes the data extracted by the attached file search means 202 together with the data obtained by the database search means 201. For example, the data displayed in the list on the search terminal 101 is processed by including at least the data such as the pamphlet and the instruction manual taken out by the attached file search means 202 individually. The display processing means 204 sends the processed data as display data that can be displayed on the display means (display such as LED) of the search terminal 101.

In the above configuration, the first crawler collecting means 503 can perform processing at an arbitrary time. For example, processing can be performed at 1-minute intervals during 6 hours from midnight. Further, the second crawler collecting means 504 can also be processed at an arbitrary time in the same manner as described above.

As described above, the first crawler collecting means 503 collects the first index table in the first index file 501, and the second crawler collecting means 504 collects the second index in the second index file 502. The table is being collected. At this time, when a keyword search request is made from the search terminal 101, the processing as shown in the flowchart shown in FIG. 6 is performed.

When a keyword search request is made from the search terminal 101, the search process is started. The first index file 501 is searched based on the keyword given from the search terminal 101, the "product number" which is the primary key information in the corresponding first index table is acquired, and the internal memory table A (FIG. 1). ) (S11).

Next, the second index file 502 is searched based on the keyword given from the search terminal 101, the "product number_serial number" which is the identification information in the corresponding second index table is acquired, and the "product number_serial number" is acquired. Only the "number" part is stored in the internal memory table B (FIG. 1) (S12).

Next, the primary key information of the internal memory table A and the logical sum of the internal memory table B are created and stored in the internal memory table C (FIG. 1) (S13). In other words, if there is the same "product number", it will be one. In this way, the database 300 and the file device 400 can be ambiguously searched, and the matching primary key information (“product number”) can be obtained at high speed. Next, the database 300 is accessed using only the primary key information of the internal memory table C, and the data corresponding to the above keyword is extracted from the obtained corresponding table (S14: database search means 201).

The data extracted above is processed (S15: display data processing means 203), sent to the search terminal 101, and displayed on the display means (S16: display processing means 204).

When the display control means 206 does not have the data retrieved by the attached file search means, the display control means 206 can display the display means based only on the data retrieved by the database search means. Further, the display control means 206 displays to the display means that a search result cannot be obtained when there is no data retrieved by the database search means 201 and data retrieved by the attached file search means 202. be able to.

In the present embodiment, the process of searching the file device 400 using the identification information "product number_serial number" to obtain data is not performed, but the information in the internal memory table B is used from the file device 400. The search may be performed by the attached file search means 202, and the obtained data may be summarized as information to be processed and displayed as a list by the display data processing means 203. This makes it possible to access the file device 400 using this information and download and display the corresponding data such as pamphlets and instruction manuals.

In contrast to the above configuration, in the present embodiment, the database search means 201 does not use the primary key information that is the search result of the first index file 501, but directly without processing the primary key information acquisition means 205. The database 300 is not searched by SQL. That is, according to such a configuration, it becomes necessary to search the column items "product number", "product name", and "product name kana" by keywords from the search terminal 101 according to the conditions of front-back intermediate match. As a database system, the load is heavy and the response is poor. If the search by the search keyword spans a plurality of tables in the database 300, the SQL for directly searching the database 300 becomes very complicated. Therefore, the above configuration is also implemented in this respect as well. It can be understood that it is not adopted in the form.

As described above, according to the present embodiment, the items (columns) distributed in each table of the database are set as the search target items by the ambiguous processing without adding special / special transformation / processing to the database and the full-text search engine. It has the effect of being able to search for the target data efficiently and at high speed (referred to as the effect of this embodiment).

Although the above embodiment is a configuration when an attached file exists, it is possible to adopt a configuration in which there is no attached file and the file device 400 is not provided. In this case, in addition to the file device 400, the second crawler collecting means 504, the second index file 502, the identification information acquiring means 207, and the attached file searching means 202 are unnecessary. Even in this case, it is possible to obtain the same effect as that of the above-described embodiment.

101 Search terminal 200 Search device 201 Database search means 202 Attached file search means 203 Display data processing means 204 Display processing means 205 Main key information acquisition means 206 Display control means 207 Identification information acquisition means 300 Database 400 File device 500 Full-text search engine 501 First index file 502 Second index file 503 First crawler collecting means 504 Second crawler collecting means

Claims (14)

A database in which multiple tables of managed data of one unit are accumulated, and
All of the plurality of tables in the database are searched by the search information, all the tables which are one unit of the managed data are specified, and the unique values for all the specified tables are used as the primary key information, and this primary key is used. First crawler collection that creates a first index table to which the search target information in the content data of the corresponding table is assigned to the information as attribute information, and generates a first index file that summarizes this first index table. Means and
When a keyword to be searched is given, the main index file is searched, the first index table having the data corresponding to the keyword is detected, and the primary key information of the first index table is obtained. Key information acquisition method and
A database search means that searches the database based on the primary key information obtained by the primary key information acquisition means and extracts data corresponding to the keyword from the obtained corresponding table.
Display means for displaying information and
A data search system comprising: a display control means for displaying the data taken out by the database search means on the display means. A file device in which attachments under the data directory of the database table are stored, and
A second index table is created by searching the file device, collecting the required data of the attached file by a crawler, and assigning the required data of the attached file to the identification information with a unique value. A second crawler collection means that generates a second index file that summarizes the index table of
When a keyword to be searched is given, the second index file is searched, a second index table having data corresponding to the keyword is detected, and identification information for obtaining identification information of the second index table is obtained. Acquisition method and
It is provided with an attachment file search means for searching the file device based on the identification information obtained by the identification information acquisition means and extracting data corresponding to the keyword from the obtained corresponding attachment file.
The data search system according to claim 1, wherein the display means displays the display means based on the data taken out by the attached file search means. Wherein the display control unit, wherein, when the attached file retrieval means there is no data to fetch, according display based only on the data the database retrieval means is taken out to claim 2, characterized in that said display means Data search system. The display control means is characterized in that the display means displays that a search result cannot be obtained when there is no data retrieved by the database search means and data retrieved by the attached file search means. The data search system according to claim 2. The data retrieval system according to claim 2, wherein the first crawler collecting means and the second crawler collecting means perform a search by either morphological analysis or N-Gram. The data retrieval system according to claim 1, wherein the first crawler collecting means performs processing at an arbitrary time. The data retrieval system according to claim 2, wherein the first crawler collecting means and the second crawler collecting means perform processing at an arbitrary time. A computer with a data search system that searches for data in a database in which multiple tables of managed data per unit are stored.
All of the plurality of tables in the database are searched by the search information, all the tables which are one unit of the managed data are specified, and the unique values for all the specified tables are used as the primary key information, and this primary key is used. First crawler collection that creates a first index table to which the search target information in the content data of the corresponding table is assigned to the information as attribute information, and generates a first index file that summarizes this first index table. means,
When a keyword to be searched is given, the main index file is searched, the first index table having the data corresponding to the keyword is detected, and the primary key information of the first index table is obtained. Key information acquisition method,
A database search means that searches the database based on the primary key information obtained by the primary key information acquisition means and extracts data corresponding to the keyword from the obtained corresponding table.
A data search program characterized in that the display means functions as a display control means for displaying data based on the data retrieved by the database search means. Further, the computer
Search the file device where the attached files under the data directory of the database table are stored, collect the required data of the attached file by crawler, and use the unique value as the identification information. A second crawler collecting means that creates a second index table to which it is assigned and generates a second index file that summarizes the second index table.
When a keyword to be searched is given, the second index file is searched, a second index table having data corresponding to the keyword is detected, and identification information for obtaining identification information of the second index table is obtained. Acquisition method,
The file device is searched based on the identification information obtained by the identification information acquisition means, and the file device is made to function as an attachment search means for extracting data corresponding to the keyword from the obtained attached file.
The data search program according to claim 8, wherein the computer is used as the display control means to cause the display means to perform a display based on the data taken out by the attached file search means. The computer is used as the display control means, and when there is no data retrieved by the attached file search means, the display means is made to function to display based only on the data retrieved by the database search means. The data search program according to claim 9. Using the computer as the display control means, the display means is indicated that no search result can be obtained when there is no data retrieved by the database search means and data retrieved by the attached file search means. The data search program according to claim 9, wherein the data search program is made to function. The data search according to claim 9, wherein the computer functions as the first crawler collecting means and the second crawler collecting means so as to perform a search by either morphological analysis or N-Gram. Program for . The data search program according to claim 8, wherein the computer functions as the first crawler collecting means so as to perform processing at an arbitrary time. The data search program according to claim 9, wherein the computer functions as the first crawler collecting means and the second crawler collecting means so as to perform processing at an arbitrary time.

Images