JP2007219627A - Retrieval method and device for large-scale information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a retrieval method and a retrieval device for large-scale information shortening a retrieval time. <P>SOLUTION: The retrieval method and retrieval device for large-scale information shortening the retrieval time can be provided by providing an indexed retrieval object database 7 in which retrieval objects are registered; a partial data storage part 3 for deletion which stores data for deletion; an indexed corresponding database 8 storing column data after deleting data for deletion from column data belonging to a retrieval object column of a record, and the identification number of the record correspondingly; an input means 1 inputting a retrieval key; a partial data deleting means 2 for deleting data for deletion from the retrieval key to create a retrieval key after deletion; a retrieval means 4 for retrieving and acquiring identification information from the indexed corresponding database 8 based on the retrieval key after deletion; and a retrieval means 4 for retrieving and acquiring the record coinciding with the acquired identification number acquired by the retrieval means 4, from the indexed retrieval object database 7. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a large-scale information search method and search apparatus that extracts information that matches an arbitrary search key from a database that holds a large amount of data.

  In the prior art, an index is created in a table holding data in order to perform a search in a short time. By searching using the index, it is possible to search in a short time.

For example, if a search is performed on a column without an index using a search key, all records are scanned and records that match the search key are extracted, which increases the search time. On the other hand, when searching for a column with an index, if the search key specified at the time of the search matches the forward match or the backward match, the index functions effectively and the search can be performed in a short time.
However, even when searching for a column with an index, if the search key specified at the time of the search does not match the front or the back, that is, if there is an intermediate match, the index does not function effectively and all records Need to be scanned, and the search time becomes longer.

As described above, when searching for a column without an index from a table holding data using the conventional search method, the search takes time compared to a column with an index, and the search is quickly performed. There is a disadvantage that records that match the key cannot be extracted.
In addition, when data with an intermediate match that does not match forward or backward is specified as a search key for a column with an index, the index is not valid and all records need to be scanned. There is a drawback that records that match can not be extracted.
In particular, when the number of stored records increases, the time required for the search increases in proportion to the amount of storage, and inconveniences such as a decrease in efficiency of work for acquiring information occur.

  Various proposals for manipulating search target character strings have been made with the aim of improving search time. For example, Japanese Patent Application Laid-Open Nos. H11-228688 and H02-209 propose a means for designating a character code to be searched, extracting a character string, and deleting a duplicate character string and a character string inclusive relation with the same character string. ing. However, since no operation is performed on the search key, if the search key does not match the index created from the character string to be searched for using the above means, the search cannot be performed quickly.

The database search time depends not only on the number of data records but also on the number of columns. In general, a search in a database is slowed by an increase in information for one record, that is, an increase in the number of columns and an increase in the capacity of the database.
Japanese Patent Laid-Open No. 9-54777 JP 11-39346 A

  The present invention has been made in view of such circumstances, and its purpose is that there is no index in the column to be searched, or even if there is an index, an intermediate match with the search key is required, so all records must be scanned. Therefore, it is an object of the present invention to provide a method for reducing the rate of search that slows down the search time and reducing the time for extracting records that match the search key.

  The present invention has been made to solve the above-described problems. The invention according to claim 1 is a large-scale information search method for searching a record in a search target storage unit based on a search key. Corresponding data storage means for storing post-deletion column data obtained by deleting deletion data from column data belonging to a search target column of the record and the identification information of the record is formed, and at the time of search, A search key after deleting the deletion data from the search key is generated, and identification information is searched and acquired from the corresponding data storage means based on the search key after deletion, and a record that matches the acquired identification information Is obtained by searching from the search target storage means.

  According to a second aspect of the present invention, in the large-scale information search method according to the first aspect, when there are a plurality of records obtained by searching the search target storage unit, the search key is extracted from the plurality of records. The large-scale information search method according to claim 1, further comprising a step of acquiring a record corresponding to.

  The invention according to claim 3 is a large-scale information retrieval device that retrieves a record in a retrieval target storage unit based on a retrieval key, and deletes deletion data from column data belonging to a retrieval target column of the record Corresponding data storage means for storing post-column data and identification information of the record in association with each other, post-deletion search key generation means for generating a post-deletion search key for deleting the deletion data from the search key, First search means for searching and acquiring identification information from the corresponding data storage means based on the search key after deletion, and a record that matches the identification number acquired by the first search means as the search target storage means And a second search means for searching and acquiring.

  According to a fourth aspect of the present invention, in the large-scale information search apparatus according to the third aspect, at least the identification information stored in the search target storage unit and the deletion stored in the corresponding data storage unit An index is created for the post-column data.

  According to a fifth aspect of the present invention, in the large-scale information search device according to the third or fourth aspect, the search performed by the first search means and the second search means uses an index. The search is based on a forward match and a backward match.

  The invention described in claim 6 is the large-scale information search apparatus according to any one of claims 3 to 5, wherein the identification information stored in the search target storage means is unique. To do.

  According to a seventh aspect of the present invention, in the large-scale information search apparatus according to any one of the third to sixth aspects, the large-scale information search apparatus is further searched by the second search means. In the case where there are a plurality of acquired records, a narrowing means for acquiring a record corresponding to the search key from the plurality of records is provided.

  The invention according to claim 8 is the large-scale information search device according to any one of claims 3 to 7, wherein the large-scale information search device further stores the record in the search target storage unit. Normal search means for searching based on the search key.

  According to a ninth aspect of the present invention, in the large-scale information search apparatus according to any one of the third to eighth aspects, the correspondence data storage means further includes column data belonging to a search target column of the record. The related information and the identification information of the record are stored in association with each other.

  According to a tenth aspect of the present invention, in the large-scale information search device according to any one of the third to ninth aspects, the large-scale information search device further inputs an input for registering a record for registration. First registration means for registering the input record and the input record and identification information for uniquely identifying the input record in the search target storage means into one record A post-deletion registration data creation means for fetching column data belonging to a search target column from the input record, and creating post-deletion registration data by deleting the deletion data from the fetched column data; and the creation And second registration means for registering the post-deletion registration data and the identification information in correspondence with the corresponding data storage means.

  According to the inventions of claims 1 and 3, in the search, since there is conventionally no index in the column to be searched, or even if there is an index, an intermediate match with the search key is required, so that it is necessary to scan all records. Even in such a case, it is not necessary to scan all the records, and it is possible to extract the records in a short time with a reduced ratio of searches for searching all the records.

  In addition, according to the inventions of claims 2 and 7, the search result is further narrowed down, so that it is possible to improve the search accuracy.

According to the invention of claim 4, since the index is created, there is an effect that the search can be performed at high speed.
Further, according to the invention of claim 5, since the search is performed by the search by the forward match or the backward match using the index, there is an effect that the search can be performed at high speed.
Further, according to the invention of claim 6, since the identification number is made unique, there is an effect that the search can be performed at high speed.
Further, according to the invention of claim 8, the normal search that has been conventionally performed can be performed, and the compatibility with the conventional search method can be maintained.
According to the ninth aspect of the invention, there is an effect that a search can be performed using an abbreviation or a common name relating to information to be searched.
Further, according to the invention of claim 10, there is an effect that data can be added to the database according to the present invention.

First, points of the embodiment of the present invention will be described.
In the embodiment, a correspondence database of identification numbers for identifying columns and records from which some data has been deleted is formed separately from the existing database. Using the corresponding database, create an index on the column from which the partial data has been deleted, and use the search key and the corresponding database from which the partial data has been deleted from the search key specified at the time of the search. The most important feature is to search for.
When searching with the search condition where the index is not valid in the existing database, extract the identification number that matches the corresponding database constructed by this means and the search key with some data deleted, and then match the extracted identification number By extracting records from an existing database, candidate records that meet the search conditions are extracted.

  Therefore, the database search is different from the search of the existing database only, and processing for searching the corresponding database is added. However, when the number of records is large, by using the search by the correspondence database, the identification number of the record corresponding to the search can be quickly specified, and the entire search time can be shortened.

  Since this means performs a search by deleting a part of the search key, a record that does not match the first search key before the deletion may be extracted. Therefore, means for re-searching with the first search key is provided for the record extracted using this means. This makes it possible to achieve both reduction in search time and improvement in search accuracy.

As described above in the embodiment of the present invention, in the present invention, an index using an “index” is important. The index will be briefly described below.
An index indicates which row (record) is located in order to improve the search speed of data. When searching for data, it is more efficient to use the index to find the location of the desired row and then read the data in that row, rather than going through every row until the desired data is found. Yes, it is a method often used in databases. In particular, a large table can be searched at high speed by using an index.

The following is a simple example of an index.
For example, it is assumed that the registered data includes four data of “Sato”, “Suzuki”, “Takahashi”, and “Watanabe”. It is assumed that a telephone number and an address are associated with each name. Here, it is assumed that “Watanabe” is searched.
If the search is performed in order, since there are four registered data, the search is completed up to four times.
Here, a search using an index is considered. Assume that there are two indexes, for example, “Sasu” corresponding to “Sato” or “Suzuki” and “Tawa” corresponding to “Takahashi” or “Watanabe”. Then, at the time of search, since “Watanabe” starts with “Wa”, the search is performed using the index “Tawa”. Next, in “Tawa”, when “Watanabe” corresponds, a telephone number or an address relating to “Watanabe” is searched.
Here, it should be noted that a search using an index always ends with two searches.

Since the above is a small number of data, even when the index is not used, the search is completed by 1 to 4 searches, and even when the index is used, the search is completed by 2 searches. However, as data increases, the effect of using an index becomes more prominent.
When there are n pieces of data, if the search is performed in order, it is necessary to search for n orders. In the case of using an index, when the amount of data increases, the index is also used in a plurality of stages. For example, as illustrated, when an index having two branches is used in multiple stages (hierarchically), the number of searches is Log2 (n). For example, in the case illustrated above, since the number of data n = 4, Log2 (4) = 2.

In creating a more general index, the index is created by various methods in consideration of the number of branches and the balance of branches. Therefore, in general, in the search of the number of data n, the order n is in the sequential search, and the search using the index is the order Log (n). Thus, the search using the index is very fast compared to the sequential search. The effect becomes more prominent as the number of data n increases.
In the present invention, the point is how to create an index effective for high-speed search in this way for existing data.

Hereinafter, a first embodiment of the present invention will be described in detail with reference to the drawings.
FIG. 1 is a block diagram showing a first embodiment of a search device according to the present invention. The search apparatus according to the first embodiment includes an input unit 1, a partial data deletion unit 2, a deletion partial data storage unit 3, a search unit 4, a registration unit 5, a search information database 6, and an output unit 9. It is configured. The search information database 6 includes an indexed search target database 7 and an indexed correspondence database 8. The indexed search target database 7 and the indexed correspondence database 8 are associated by an identification number. Preferably, the identification number of the indexed search target database 7 is unique, and the identification number of the indexed search target database 7 is uniquely specified from the identification number of the indexed correspondence database 8.

  Next, as a first embodiment, a case where a search for a corporate name is accelerated in a database that manages customer information will be described in detail below.

  In the deletion partial data storage unit 3, deletion partial data to be deleted from the search key input from the input unit 1 is registered. For example, “corporation”, “(stock)”, “limited company”, “(existing)”, “foundation”, “(goods)”, etc., representing the legal entity are registered.

  At the time of search, the input unit 1 performs a process of inputting a search key input by the operator, and sends the input search key to the partial data deletion unit 2 and the search unit 4. Further, at the time of data registration, the registration data input by the operator is input, and the input registration data is sent to the partial data deletion means 2 and the registration means 5.

The partial data deleting means 2 uses the search key sent from the input means 1 at the time of searching, and uses partial data for deletion registered in the partial data storage section 3 for deletion, for example, “stock” Delete "Company", "Corporation", etc. In this way, a new search key is created by deleting the deletion partial data from the search key. Hereinafter, this new search key will be referred to as a partial deletion search key. Next, the partial data deletion unit 2 sends a partial deletion search key to the registration unit 5.
Further, the partial data deletion means 2 extracts only the part corresponding to the column (field) to be searched from the registration data sent from the input means 1 at the time of registration, and from the extracted registration data, The deletion partial data registered in the deletion partial data storage unit 3, such as “corporation” or “(share)” representing a legal entity, is deleted. In this way, new registration data is created by deleting the deletion partial data from the registration data. Hereinafter, this new registration data will be referred to as “partially deleted registration data”. Next, the partial data deletion unit 2 sends the partial deletion registration data to the registration unit 5.

  The search means 4 searches the search information database 6 using the search key sent from the input means 1 and the partial deletion search key sent from the partial data deletion means 2.

  The registration unit 5 registers the registration data sent from the input unit 1 and the partial deletion registration data sent from the partial data deletion unit 2 in the search information database 6.

  The search information database 6 performs a search using the search key and the partial deletion search key input from the search unit 4 and outputs the search result to the output unit 9. Also, registration data and partial deletion registration data input from the registration means 5 are registered. Details of the search information database 6 will be described later in detail with reference to FIGS. 2, 3, and 4.

  The output means 9 displays the search result from the search information database 6 to the operator. Alternatively, the registration result is displayed to the operator.

  Next, details of the search information database 6 will be described with reference to FIG. As described above with reference to FIG. 1, the search information database 6 includes an indexed search target database 7 and an indexed correspondence database 8. In FIG. 2, the configuration of the search information database 6 that is assumed to manage customer information data is illustrated as an example.

The indexed search target database 7 includes a customer information table 21 and its index 22.
In the customer information table 21, customer information such as a corporate name, an address, a telephone number, and the like related to a customer and an identification number are registered as one record. An index 22 is created for the identification number column of the customer information table 21. Since the index 22 is created in the identification number column, when a search for a specific identification number is performed in the identification number column of the customer information table 21, a record corresponding to the specific identification number is searched in a short time.
The identification number is a unique key, for example, and the customer information record can be uniquely specified by specifying the identification number.

In the customer information table 21 of FIG. 2, as an example of a customer record, an identification number “50001”, a corporate name “Nippon Nippon Telegraph and Telephone Corporation”, an address “Osaka ...”, and a phone number “06 ...” Etc. are registered. In this case, “50000 series” is created and registered as the index 22 for the identification number “50001”.
Such an indexed search target database 7 has the same configuration as an existing customer information database that has been conventionally used.

Next, there is an indexed correspondence database 8 corresponding to the indexed search target database 7. The indexed correspondence database 8 includes a correspondence table 23 and its index 24.
For example, it is assumed that a legal personality is designated as partial data to be deleted (hereinafter, partial data for deletion). The legal entity is, for example, “corporation”, “(stock)”, “limited company”, “(present)”, “foundation”, “(goods)”, or the like. The partial data for deletion is designated in the partial deletion data storage unit 3 by, for example, “corporation”, “(share)”, “limited company”, “(present)”, “foundation”, “( Etc.) ”is registered.

  Since the partial data for deletion is designated, the correspondence table 23 includes data obtained by deleting the partial data for deletion from the corporate name that is the search target data registered in the customer information table 21. It is registered as a name excerpt. For example, the corporate name “corporation” registered as partial data for deletion is deleted from the corporate name “Nippon Telegraph and Telephone Corporation” in the customer information table 21, and “Nishi Nippon Telegraph and Telephone” is extracted from the corporate name in the correspondence table 23. Is registered. In addition, the corporate name excerpt “Nippon Nippon Telegraph and Telephone” and the identification number “50001” for identifying the record in the customer information table 21 are registered in the correspondence table 23 as a pair.

That is, in the correspondence table 23, a deletion part such as “stock” or “stock” registered in the deletion part deletion data storage unit 3 from the data in the search target column of the customer information table 21. Data from which data has been deleted (hereinafter, “partially deleted data”) and an identification number that can uniquely identify a record in the customer information table 21 are registered as a pair.
Further, in the indexed correspondence database 8, an index 24 is created for the corporate name excerpt (partially deleted data) in the correspondence table 23.

  In the indexed search target database 7 and the indexed correspondence database 8, the creation of an index for a specific column is performed by the function of the database. The creation of an index and the automatic update of the index for a specific column of such a specific database are provided by normal database functions.

  Next, the operator searches using the input means 1 (A1). For example, "Nippon Nippon Telegraph and Telephone Corporation" is entered as a search key (A2). Next, from the input search key, the search for the corporate name is started using the indexed correspondence database 8 and the indexed search target database 7 (A3). Details of processing performed at the time of this search will be described in detail with reference to FIG.

  Processing performed in the search in the first embodiment will be described with reference to the flowchart of FIG. First, the operator uses the input unit 1 to input a search key and starts a search (step S31). For example, “Nippon Nippon Telegraph and Telephone Corporation” is input as a search key, and the search is started. Next, the partial data deletion means 2 deletes the partial deletion data registered in the deletion partial data storage unit 3 from the input search key (step S32), whereby the partial deletion search is performed. Get the key. For example, from the search key “Nippon Nippon Telegraph and Telephone”, the data for partial deletion “Corporation” representing the legal entity is deleted, and “Nippon Telegraph and Telephone” is acquired as the partial deletion search key. Next, using the acquired partial deletion search key “Nippon Nippon Telegraph and Telephone”, the search means 4 acquires an identification number that matches the condition from the indexed correspondence database 8 using the index 24 (step S33). . For example, the identification number “50001” corresponding to “Nippon Nippon Telegraph and Telephone” is acquired from the indexed correspondence database 8 using the index “west” of the index 24. Next, using the acquired identification number (as a search key), a record that matches the identification number is acquired from the indexed search target database 7 using the index 22 (step S34). For example, by using “50000 series” of the index 22, as the record whose identification number matches “50001” from the indexed search target database 7, the corporate name “Nippon Nippon Telegraph and Telephone Corporation”, the address “Osaka ...” The telephone number “06-...” Is acquired. Next, the obtained list of records is displayed to the operator using the output means 9 (step S35). For example, the corporate name “Nippon Nippon Telegraph and Telephone Corporation”, the address “Osaka ...”, the telephone number “06-...”, Etc. are displayed on the output means 9.

  The search using the index in step S33 and the search using the index in step S34 can use forward matching and / or backward matching.

  In the first embodiment, when there is only one record corresponding to the search, only one record is output. When a plurality of records correspond, the corresponding records are displayed as a list. Further processing when a plurality of records are applicable will be described in the second embodiment.

  As described above, in the first embodiment, the indexed correspondence database 8 is registered with the corporate name excerpt obtained by deleting the data for partial deletion such as the corporate case from the corporate name of the indexed search target database 7. In addition, an index is created for this corporate name excerpt. Furthermore, in the case of a search, the partial deletion data such as a legal case is deleted from the input search key, a partial deletion search key is obtained, and the index-corresponding database 8 is stored using this partial deletion search key. A search is performed, and as a search result, first, an identification number for specifying a record of a specific customer in the indexed search target database 7 is obtained. Hereinafter, this search is referred to as a primary search. Such a primary search and the configuration for the primary search have the following advantages.

  In the prior art, an index may be created for the corporate name column of the customer information table 21 in the indexed search target database 7. However, even if an index is created for a corporate name, the index may not work effectively. One of the reasons is that the corporate name has a legal character. In general, in order to create an effective index, it is important that the amount of data to be indexed is small and there is little duplication. However, if the corporate name such as a corporation is included in the data in a large amount like the corporate name and there are duplicates, the index may not be created effectively. As a result, when searching for a corporate name using the corporate name index, the search may not be fast.

  Compared to creating an index on the corporate name column of the conventional indexed search target database 7, an index is created for the corporate name excerpt of the indexed correspondence database 8 of the first embodiment. Has the following advantages. First, since the data for partial deletion is deleted, there is an advantage that the amount of data to be indexed is reduced. Next, there is an advantage of reducing duplicate data such as corporations. With these advantages, an index effective for high-speed search is created for the corporate name excerpt column of the indexed correspondence database 8. As a result, there is an advantage that the search for the corporate name excerpt column in the indexed correspondence database 8 becomes faster.

  In general, it is known that in a database to be searched, it is faster to search a database having a small number of columns if the number of records is the same. Also in this respect, there is an advantage that the search in the indexed correspondence database 8 having only two columns of the corporate name excerpt and the identification number is faster than the search in the conventional indexed search target database 7 having a large number of columns. is there.

  Also, for example, when an operator enters a corporate name and searches, even if it is the same legal entity, when entering a corporate case at the beginning or at the end, when entering an abbreviated corporate case, There are cases where input is made with labor saving. For example, “Iroha Co., Ltd.”, “Iroha Co., Ltd.”, “(Iroha Co., Ltd.)”, “Iroha”, etc. may be entered and searched.

Considering the input of the search key by the operator as described above, the search in the case where an index is created for the corporation name of the conventional indexed search target database 7 has the following problems. When the corporate name registered in the customer information table 21 is registered with a corporate case at the beginning, if a search is performed using a search key with the corporate case at the end, the search is not matched because the search condition is not met. For example, if the registered corporate name is “Iroha Corporation” and the search key used for the search is “Iroha Corporation”, they are not matched and are not searched.
Further, in this case, even if the operator manually inputs the search key “Iroha” and performs a search with a forward match, the registered corporation name does not start with a forward match because it starts with “corporation”. In this case, since partial matches are made, it is necessary to search all records. Eventually, the index will not be valid and the search will be slow.
On the other hand, if a backward match is performed instead of a forward match in the search, the index works effectively and a high-speed search may be possible. In the above example, if a backward match search is performed on the registered data “Iroha Co., Ltd.” with the search key “Iroha”, the index works effectively and a high-speed search is possible. However, when the data registered in the database is not “Iroha Co., Ltd.” but “Iroha Co., Ltd.”, conversely, the search using the index is effective only by the forward matching. In any case, the conventional search has a problem that the index is not always effective depending on how the corporate name of the corporate name is registered and how to search.

On the other hand, in the primary search according to the first embodiment, the search matches not the search key but the data that matches the partial deletion search key by searching for the corporate name in the indexed correspondence database 8, so that the search matches forward or backward The index works effectively in both conditions. Or, the probability that the index works effectively increases. Therefore, in the primary search, an identification number that matches the partial deletion search key can be quickly extracted.
For example, even if “Iroha Co., Ltd.” or “Iroha Co., Ltd.” is registered in the indexed search target database 7, the corporate name excerpt “Iroha” is registered in the indexed correspondence database 8. Even if the operator performs a search using a search key such as “Iroha Co., Ltd.” or “Iroha Co., Ltd.”, the partial deletion search key is “Iroha”. Therefore, when a search is performed using the partial deletion search key in the indexed correspondence database 8, the index functions effectively under either the forward match or backward match search condition, and high-speed search is possible.

  Further, by using the indexed correspondence database 8, even if there is no forward match or backward match, an intermediate match, and a search of all records is required, it is more preferable than a search in the indexed search target database 7 having a large number of columns. If the search is performed in the indexed correspondence database 8 that has only two columns of the corporate name excerpt and the identification number, the search is performed faster. In addition, regarding the amount of data to be searched, the data amount of the corporate name excerpt registered in the indexed correspondence database 8 is smaller than the data amount of the corporate name in the indexed search target database 7. The search is made faster.

Next, using the identification number acquired by the primary search, a search is performed from the indexed search target database 7 to acquire a record that matches the identification number. Such a search is hereinafter referred to as a secondary search.
In the indexed search target database 7, an index 22 is created for the identification number of the customer information table 21. The identification number is, for example, a unique natural number, has no duplication, and has a small amount of data. Therefore, the index 22 for the identification number is effectively created in terms of high-speed search. Therefore, the search for the identification number in the indexed search target database 7 is performed at high speed because the index 22 functions effectively. Therefore, this secondary search is performed at high speed.

As described above, in the primary search, the index is effectively created by deleting the partial deletion data in the column data to be searched, and the partial deletion data is deleted from the input search key. The search is performed by the copy deletion search key, and since a small database is used, a high-speed search is possible. Further, since the secondary search is a search based on an identification number with an index, generally a high-speed search using the index is possible. As described above, both the primary search and the secondary search are performed at high speed. For this reason, two searches of the primary search and the secondary search are required, but the entire search can be performed at high speed.
In addition, the effect that the high-speed search can be performed according to the first embodiment becomes remarkable as the amount of data to be searched increases. The reason why the above effects can be obtained is because the index is effectively created by deleting some data for deletion from the data to be searched, and the search that can utilize the index is performed. is there.

  As described above, according to the first embodiment, for example, in a case where a search is performed by switching the legal case before and after, the conventional technique requires scanning of all records without a front match or a back match. In addition to the indexed search target database 7, the indexed correspondence database 8 is created with the data from which the legal case has been deleted, and the indexed correspondence database 8 is used to search. You can search. For this reason, there is an advantage of quickly extracting records that match the search key without scanning all records.

  Next, a procedure for registering data will be described. In the case of performing conventional customer information management, data registration is only registration in the customer information table 21. On the other hand, in the first embodiment, when data is registered, it is registered in the correspondence table 23 simultaneously with the registration in the customer information table 21. The registration procedure will be described in detail with reference to the flowchart of FIG.

First, search target data to be registered by an operator is input from the input means 1 (step S41). Next, the registration means 5 registers the input data as a record in the customer information table 21 that is a search target table (step S42). At the same time, for example, the registration unit 5 creates an identification number for the input data, and registers the data and the identification number to be input to the search target table as one record. Alternatively, for example, an identification number for the input data may be created by the function of the indexed search target database 7, and the input data and the created identification number may be registered as one record. For example, the identification number created and registered here is created and registered so as to be a unique key that can uniquely identify the record.
Thereafter, in the indexed search target database 7, an index for the identification number is automatically created. This is automatically done because it is specified by the general database function that an index is created for the identification number. The above registration of customer information data is the same as conventional registration of customer information data.

Next, the part (field) which becomes the search target column of the input data by the partial data deleting means 2 matches the partial deletion data registered in the partial deletion data storage unit 3. Is determined (step S43). If not included, the part that is the search target column of the input data (this is partly deleted data) and the previously created identification number are registered as one record in the correspondence table 23 (step S45). On the other hand, if it is included, the matching partial deletion data is deleted from the portion that becomes the search target column of the input data (step S44). Next, the data obtained by deleting the data for partial deletion that matches from the part that becomes the search target column of the input data (this is referred to as partial deletion data) and the previously created identification number as one record And registered in the correspondence table 23 (step S45).
Thereafter, in the indexed correspondence database 8, an index is automatically created for the corporate name excerpt that is a column to be searched. This is automatically done because it is specified by the general database function that an index is created for the corporate name excerpt.

  As described above, the input data and the identification number are registered as one record in the customer information table 21 (indexed search target database 7), and the partially deleted data and the identification number are recorded as one record in the correspondence table. 23 (indexed correspondence database 8). After that, each index is created by the database function.

  In the above description of registration, data was registered in two databases, and an index was created after each registration. However, in order to ensure consistency in registering data in the two databases, transaction processing is performed in registering data in the two databases, and after the two registrations are successfully performed (after committing), the index May be updated. As described above, normal arbitrary processing in the database may be further used.

  Next, a second embodiment for further narrowing down records when there are a plurality of retrieved records will be described. In the second embodiment, a search result is obtained by performing a search or extraction again with the original search key that has not deleted the partially deleted data for the record that has been searched with the partially deleted search key. Can be narrowed down. The configuration and operation will be described in detail below.

FIG. 5 is a diagram illustrating the configuration of the search device according to the second embodiment. In this configuration, the narrowing means 10 is further added to the configuration in the first embodiment.
The narrowing means 10 performs a narrowing process on the search result from the search information storage unit 6 using the partial deletion key, using the search key input from the input means 1. Next, the narrowed search result is output to the output means 9.
Other configurations are the same as those of the first embodiment.

First, a case where a search using a partial deletion search key is performed as in the first embodiment, a plurality of records may be searched, will be described with reference to FIG. In FIG. 6, as in the customer information management in the first embodiment, there are a customer information table 61 and a correspondence table 62 for the indexed search target database 7 and its indexed correspondence database 8.
In the customer information table 61, as an identification number and a corporate name as an example, “1001, ABC Corporation”, “1002, ABC Corporation”, “60001, Limited Company ABC”, “60002, Foundation ABCD”, etc. Is registered. In addition, an address and a telephone number for each corporate name are registered. In the deletion partial data storage unit 3, “Co., Ltd.”, “(share)”, “limited company”, “(present)”, “foundation”, “(goods)”, etc. are registered. To do. In this case, in the correspondence table 62, a corporate name excerpt is created from the customer information table 61 and the partial data storage unit 3 for deletion, and the corporate name excerpt and the identification number are “ABC, 1001”, “ABC, 10002 "," ABC, 60001 "," ABCD, 60002 ".
In the above description, indexes and tables in the database are omitted for the sake of simplicity. The detailed configuration of the indexes and tables in the database is the same as in the first embodiment.

Here, “ABC Inc.” is input as a search key, and a search is performed using the search key “ABC Inc.” (B1). In the partial data storage unit 3 for deletion, since “corporation” is registered as partial deletion data, “corporation” is deleted from the search key “ABC” (B1). “ABC” is acquired as a deletion search key (B2). Next, a primary search is performed from the indexed correspondence database 8 using the partial deletion search key “ABC” (B2).
The results of the primary search from the indexed correspondence database 8 using the partial deletion search key “ABC” (B2) are the identification numbers “1001”, “1002”, “60001”, and “60002” (B3). The search here is assumed to be a forward match. Next, the result of the secondary search from the indexed search target database 7 by the identification number (B3) of the primary search result is “1001, ABC Corporation, address,..., Telephone number. 1002, ABC Co., Ltd., Address ..., Phone Number ... "," 60001, ABC Company, Address ... Phone Number ... "," 60002, ABCD, Address ... " Phone number ... "(B4).

  As described above, since a search is performed using the partial deletion search key, a plurality of search targets may be searched. In addition, since the search key first input as the object of the search target is “ABC Inc.”, a search key that does not necessarily match the search key may be searched.

  When a plurality of records have been searched in this way, it is possible to extract a record intended for search by further narrowing down the search results. This is performed by further performing a narrowing process for searching or extracting a column as a search target by using the search key by the narrowing means 10 on the search result.

  For example, the narrowing means 10 for the corporation name “ABC Co., Ltd.”, “ABC Co., Ltd.”, “ABC Ltd.”, and “ABC ABC Corporation” of the search results, which are the results of the above secondary search, By searching or extracting a search key that matches the search key “ABC”, only “1001, ABC, Inc., address,..., Telephone number. Next, this search result is output to the output means 9.

Next, the operation of the narrowing-down process performed in the second embodiment will be described using the flowchart of FIG.
The flowchart of FIG. 7 is obtained by adding a process when there are a large number of extracted record candidates in the flowchart of the search process performed in the first embodiment of FIG. Steps S71 to S74 in the flowchart of FIG. 7 are the same as the processes performed in the search of the first embodiment, and are the same as steps S31 to S34 in the flowchart of FIG. Below, a different part from the search process in 1st embodiment is demonstrated.

  First, a record that matches the partially deleted data is acquired by the processing up to step S74. That is, the result of the secondary search is acquired. Next, it is determined whether there are two or more acquired records (step S75). If the acquired record is less than 2, that is, 1 or 0, the acquired record is displayed (step S76). This is the same as the processing performed in the first embodiment. On the other hand, when there are two or more acquired records, a narrowing process using a search key is further performed on the list of acquired records (step S77). Next, a list of records obtained and narrowed down is displayed (step S78).

  As described above, the narrowing-down process performed in step S77 described above is to extract only the part that is the search target of the secondary search result that matches the first input search key. Is made by Alternatively, the result of the secondary search may be temporarily stored in another table, and the search may be performed from there using a search key. The specific extraction or search is arbitrary, and it is important to perform the extraction or search again with the search key input first.

  As described above, by performing the narrowing-down process in step S77, even when a plurality of records correspond to the search using the partial deletion search key, only the search key that matches the search key first input by the operator is acquired and displayed. Is done.

As additional processing, further processing may be performed depending on the number of results narrowed down in step S77. For example, if there are a plurality of search results in step S74 and the number of results narrowed down in step S77 is 0, it is also possible to display the search result records in step S74 before narrowing down as a list.
In the second embodiment, other processes such as registration are performed in the same manner as in the first embodiment.

Next, another example of the first or second embodiment for a system that manages a large number of confectionery in the indexed search target database 7 will be described with reference to FIG.
A place name is specified as the data for partial deletion to be deleted, the product name region deletion in which the data for partial deletion is deleted from the product name, and the identification number identifying the record in the product management table as a pair, the indexed correspondence database 8 create.
When searching for confectionery with a place name at the beginning of a product name, when the operator manually designates the product name as a search key, a search may be performed using different place names under a forward matching condition. In this case, the data registered in the indexed search target database 7 does not match the forward matching condition, and a record cannot be extracted. However, by performing a search using the search key from which the place name has been deleted (partially deleted search key) and the indexed correspondence database 8, it is possible to extract a record that matches the search key from which the place name has been deleted. Next, by searching the indexed search target database 7 using the identification number included in the extracted record as a search key, it is possible to quickly extract record candidates using the index effectively.

  As described above, in the embodiment according to the present invention, the target database is not limited to the corporate name, and can be applied to confectionery product management and other various databases.

Next, a third embodiment in which a plurality of records in the indexed correspondence database 8 correspond to one record in the indexed search target database 7 will be described.
For example, an indexed search target database 7 (FIG. 2) that manages customer information is assumed as in the first embodiment. In the first embodiment, only “Nippon Nippon Telegraph and Telephone, 50001” is registered in the indexed correspondence database 8 with respect to “50001, West Nippon Telegraph and Telephone Corporation” in the customer information database. However, in the indexed correspondence database 8, it is possible to register a plurality of corresponding records for “50001, Nippon Telegraph and Telephone Corporation” in the customer information database.

  For example, in addition to “Nippon Telegraph and Telephone, 50001”, the indexed correspondence database 8 can have a plurality of records for “West Japan NTT, 50001”, “West NTT, 50001”, and the identification number 50001. is there. In this way, in the indexed correspondence database 8, not only partially deleted data but also alias names such as abbreviations and common names can be registered.

  The search in the third embodiment is the same as the search performed in the first embodiment. For example, the partial deletion data of the legal personality is deleted for the search key “West Japan NTT” which is an abbreviation. In this case, since there is no matching corporate case, nothing is deleted from the search key, and “West Japan NTT” is acquired as the partial delete search key. With this partial deletion search key, a search is performed in the indexed correspondence database 8 to obtain an identification number 50001. After that, the indexed search target database 7 is searched using the identification number 50001 as a search key. As described above, even when a plurality of records corresponding to one record of the indexed search target database 7 are registered in the indexed correspondence database 8, the configuration and processing are the same as in the first embodiment. is there. Further, as in the second embodiment, it is possible to perform a narrowing process.

As described above, records having the same identification number can be registered in the indexed correspondence database 8. On the other hand, in the indexed search target database 7, the identification number is unique, and it is important that the record can be uniquely identified by the identification number.
Since the identification number is configured in this way, the search from the identification number of the indexed correspondence database 8 to the indexed search target database 7 is uniquely performed.

  In the third embodiment, the registration method in the indexed correspondence database 8 is an arbitrary design. For example, when the operator selects an arbitrary customer record from the indexed search target database 7 and the operator inputs an abbreviation or common name for the customer and selects to register it, the input abbreviation or common name and index are entered. The identification number of the customer in the indexed search target database 7 is registered as one record in the indexed correspondence database 8.

  For example, in the above-described first embodiment, the search using the correspondence database is performed only for the corporate name as the column for performing the search. However, the search can also be performed in other columns. For the address column, telephone number column, etc., as with the corporate name column, a corresponding database is created for each column, and the corresponding database for each column is created. It is possible to obtain an identification number by performing a primary search, and then perform a secondary search by the identification number.

In the case of such a configuration, in the deletion partial data 3, for each column to be searched, a corporate case such as “corporation”, which is data to be deleted, a local telephone number such as “03-”, etc. sign up. A correspondence database for each column is also created in the indexed correspondence database 8. In the correspondence database, deletion data for each column is deleted from the column data to be searched and registered.
Also, at the time of search or registration, the part deletion data specified for each column to be searched is deleted for the search key, a part deletion search key is created, and part deletion search is performed. Search or register by key.
As described above, it is also possible to configure such that a plurality of columns can be searched and registered.

The first to third embodiments of the present invention can be added to an existing information management system.
For example, in the conventional customer information management system, the configuration is completed by the configuration of the input unit 1, the search unit 4, the registration unit 5, the indexed search target database 7, and the output unit 9 in the first embodiment. It was. That is, the customer information is searched and registered only in the indexed search target database 7, that is, the indexed search target database 7. By adding the partial data deletion means 2, the deletion partial data storage section 3, and the indexed correspondence database 8 to the conventional configuration and changing the processing performed in the search means 4 and the registration means 5, the first This is an embodiment. Therefore, by adding a small number of configurations to an existing customer information management system that is already in operation, it is easy to operate the configuration as in the first embodiment. In addition, it is easy to configure from the first to the third by adding a configuration to a system using an existing database.

For the search in the first to third embodiments, the search means 4 searches the indexed correspondence database 8 using the partial deletion search key, obtains the identification number, and stores the index search target data by the identification number. A search for part 7 was performed. However, the search means 4 can also search the indexed search target database 7 directly from the search key. This is a conventional search method. As described above, the search method according to the embodiment of the present invention can coexist with the conventional search method.
Moreover, it is also possible to attach an index to an arbitrary column to be searched in the indexed search target database 7 for the conventional search method.
As described above, the first to third embodiments can be easily expanded while leaving the existing existing database and the search method thereof.

In the embodiment, for example, the identification number of the indexed search target database 7 is a unique key. However, the identification number is not limited to a unique key, and may be a primary key. Further, the identification number may be a unique key or a primary key using a plurality of columns. In this case, a plurality of columns are also used for identification information in the correspondence database. An important point as an embodiment in the present invention is that a corresponding record can be uniquely identified by an identification number.
Although the identification number has been described as a unique natural number, the identification number is not limited to a natural number, and may be a character other than a number. Even in this case, there is a demand for characters that can uniquely identify each record.

  Further, for example, the indexed search target database 7 and the indexed correspondence database 8 may be configured as one database or may be configured as separate databases. Alternatively, the two databases may be separate databases via a network.

  Moreover, the search means 4 and the registration means 5 can also be made into a part of function of the indexed search object database 7 and the indexed correspondence database 8 as the function of the database.

  Moreover, each function in the first to third embodiments can be used independently, or can be used in any combination.

  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 a design and the like within the scope not departing from the gist of the present invention.

  The present invention is most suitable for use in a database search method that holds a large amount of data.

It is a block diagram which shows the structure of the search device which adapted the search device of the large-scale information by 1st Embodiment of this invention. It is a figure which shows an example of the memory content of the indexed search object database 7 and the indexed correspondence database 8 in the search device. It is a flowchart which shows the procedure of the search of data in the search device. It is a flowchart which shows the procedure of registration of data in the search device. It is a block diagram which shows the structure of the search device which adapted the search device of the large-scale information by 2nd embodiment of this invention. It is a block diagram which shows the example which a some search result produces in the same search device. It is a flowchart which shows the procedure of the search of data in the search device. It is a figure which shows another example of the memory content of the indexed search object database 7 and the indexed corresponding | compatible database 8 in the search device which applied the large-scale information search device by 1st or 2nd embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Input means 2 ... Partial data deletion means 3 ... Deletion partial data storage part 4 ... Search means 5 ... Registration means 6 ... Search information database 7 ... Retrieval database with index 8 ... Corresponding database with index 9 ... Output means 10 ... Narrowing means

Claims (10)

In a large-scale information search method for searching records in a search target storage unit based on a search key,
Forming correspondence data storage means for storing the post-deletion column data obtained by deleting the deletion data from the column data belonging to the search target column of the record and the identification information of the record in association with each other;
When searching,
Generate a post-deletion search key that deletes the deletion data from the search key,
Based on the search key after deletion, the identification information is retrieved from the corresponding data storage means and acquired,
A search method for large-scale information, wherein a record that matches the acquired identification information is searched and acquired from the search target storage means.   2. The method according to claim 1, further comprising a step of acquiring a record corresponding to the search key from the plurality of records when there are a plurality of records acquired by searching the search target storage unit. How to search for scale information. In a large-scale information retrieval apparatus that retrieves a record in a retrieval target storage unit based on a retrieval key,
Corresponding data storage means that stores post-deletion column data in which deletion data is deleted from column data belonging to the search target column of the record, and identification information of the record in association with each other,
A post-deletion search key generation means for generating a post-deletion search key by deleting the deletion data from the search key;
First search means for searching and acquiring identification information from the corresponding data storage means based on the post-delete search key;
A second search means for searching and acquiring a record that matches the identification number acquired by the first search means from the search target storage means;
An apparatus for retrieving large-scale information, comprising:   The index is created for at least the identification information stored in the search target storage unit and the post-deletion column data stored in the corresponding data storage unit. Large-scale information retrieval device.   5. The large-scale information according to claim 3, wherein the search performed by the first search means and the second search means is a search based on a forward match and a backward match using an index. Search device. 6. The large-scale information search apparatus according to claim 3, wherein the identification information stored in the search target storage means is unique. The large-scale information search device further comprises:
When there are a plurality of records searched and acquired by the second search means, a narrowing means for acquiring a record corresponding to the search key from the plurality of records,
The large-scale information search apparatus according to claim 3, wherein the large-scale information search apparatus includes: The large-scale information search device further comprises:
Normal search means for searching records in the search target storage means based on the search key;
The large-scale information search device according to claim 3, wherein the large-scale information search device is provided. The corresponding data storage means further comprises:
Storing information related to column data belonging to the search target column of the record in association with identification information of the record;
The large-scale information search device according to claim 3, wherein The large-scale information search device further comprises:
An input means for inputting a record for registration;
A first registration means for registering the input record and identification information for uniquely identifying the input record in the search target storage means into one record, and registering the search record in the search target storage means;
Post-deletion registration data creation means for retrieving column data belonging to a search target column from the input record, and creating post-deletion registration data by deleting the deletion data from the extracted column data;
Second registration means for registering the created post-deletion registration data and the identification information in correspondence with the corresponding data storage means;
The large-scale information search apparatus according to claim 3, wherein the large-scale information search apparatus includes:

Images