JP2011253299A - Retrieval device, retrieval method and retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the response time of retrieval without unconformity with a database.SOLUTION: Respective documents constituting response data are stored in a fragment cache storing part 111. A document acquisition part 106 acquires documents from the fragment cache storing part 111 if the documents to be acquired exist in the fragment cache storing part 111. If there is no such document, the document acquisition part 106 acquires documents from an RDBMS2. As a result, the number of times of acquisition of documents from the RDBMS2 can be minimized, thereby reducing the response time of retrieval.

Description

  The present invention relates to a technique for searching metadata.

  2. Description of the Related Art Conventionally, in an application server that retrieves metadata corresponding to a search query from a client from the Relational DataBase Management System (RDBMS) and returns the search result, the search query and the search result of the initial inquiry are stored as a cache. The response time is shortened by returning the search result from the cache when there is an inquiry of the same search query after the first time (for example, Patent Document 1).

JP 2009-175896 A

  In the application server, when the metadata is a structured document such as xml stored in a large object type, the search time on the RDBMS and the acquisition time of the large object from the RDBMS become a bottleneck. A large object is a type of data stored in an RDBMS and has a feature that data having a size exceeding the data page can be stored.

  As in the prior art, return data is cached at the time of the first search, and the return data stored in the cache is used for the subsequent inquiries, thereby shortening the response time and reducing resource consumption. Can do. However, in the conventional cache that manages the search query and the return data on a one-to-one basis, when the return data is a set of large object type data stored on the RDBMS, the return data is configured on the RDBMS. When one large object type data is updated, there is a problem that inconsistency occurs between the cached return data and the RDBMS.

  The present invention has been made in view of the above, and an object of the present invention is to shorten the response time of a search without causing inconsistency with a database when return data is a set of large object type data.

  A search device according to a first aspect of the present invention is a search device for searching for a document corresponding to a search condition from a database storing documents, receiving means for receiving a search query, and the database using the search query. Search means for obtaining a search result having a storage location information indicating the identifier of the document and the storage location of the document, and a previously acquired document, A first cache storage unit for storing an identifier, and a determination is made as to whether or not a document corresponding to the identifier of the document of the search result exists in the first cache storage unit, and the document exists in the first cache storage unit In the case, from the first cuche storage means, An acquisition means for acquiring the document from the database based on the storage location information of the search result, and the first cache storage of the document acquired from the database together with an identifier of the document. Storage means for storing in the means, generation means for combining one or more documents acquired by the acquisition means to generate response data, and return means for returning the response data .

  A search method according to a second aspect of the present invention is a search method in which a search device searches for a document corresponding to a search condition from a database storing documents, the step of receiving a search query, and using the search query Searching the database, obtaining a search result having storage location information indicating the identifier of the document and the storage location of the document with respect to a document corresponding to the search condition of the search query, and the identifier of the document of the search result It is determined whether or not a document corresponding to the above-mentioned document exists in the first cache storage unit that stores the previously acquired document and the identifier of the document, and if the document exists in the first cache storage unit, The document is saved from one cuche storage means. And if not present, obtaining the document from the database based on the storage location information of the search result, and storing the document obtained from the database together with an identifier of the document in the first cache storage means A step of combining the one or more documents acquired by the acquiring unit to generate response data, and a step of returning the response data.

  A search program according to a third aspect of the present invention is a search program for causing a computer to execute the search method.

  According to the present invention, when the return data is a set of large object type data, the search response time can be shortened without causing inconsistency with the database.

It is a functional block diagram which shows the structure of the search device in this Embodiment. It is the schematic explaining a remponce cache and a fragment cache. It is a flowchart which shows the flow of a process of the search device in this Embodiment. It is a flowchart which shows the flow of the process which acquires a document.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a functional block diagram showing the configuration of the search device 1 according to the present embodiment. The search device 1 shown in the figure is a device that receives a search query from the terminal 3, searches the RDBMS 2 for program metadata of a broadcast program and video-on-demand content, and returns it to the terminal 3. The search query is an information search request from the terminal 3 to the search device 1 and is a combination of search conditions such as title, genre, date and time, using logical operators and modifiers. The terminal 3 transmits a search query to the search device 1 and receives a search result for the search query as response data. For example, the terminal 3 includes a receiving terminal in the Internet Protocol Television (IPTV) service.

  First, the configuration of the RDBMS 2 will be described. The RDBMS 2 includes a plurality of attribute tables 201, a document table 202, a primary search result table 203, and a cache management table 204. The attribute table 201 is an expansion of a document in which program metadata is described, and includes information (program metadata) related to a program such as a content ID (crid), a title, an outline, and performers. A document is a structured document described in XML or the like, and an attribute is a component of a document described by name / value pairs. The document table 202 stores a fragment object (fragment) obtained by dividing a fragment ID (fid), fragment version (fver), content ID (crid), and program metadata (fragment) into return units. The primary search result table 203 stores a hash value generated from the search query and a list (content list) of content IDs having program metadata corresponding to the search condition. The cache management table 204 stores a hash value, a primary search result stored in the primary search result table 203, and an expiration date of a response cache described later.

  Next, the configuration of the search device 1 will be described. As shown in FIG. 1, the search device 1 includes a query receiving unit 101, a cache expiration date determination unit 102, a syntax analysis unit 103, an SQL generation unit 104, a search execution unit 105, a document acquisition unit 106, a response data generation unit 107, A response data return unit 108, a response cache storage unit 110, and a fragment cache storage unit 111 are provided. Each unit included in the search device 1 may be configured by a computer including an arithmetic processing device, a storage device, and the like, and the processing of each unit may be executed by a program. This program is stored in a storage device included in the search device 1, and can be recorded on a recording medium such as a magnetic disk, an optical disk, or a semiconductor memory, or provided through a network.

  The query receiving unit 101 receives a search query from the terminal 3.

  The cache expiration date determination unit 102 calculates the hash value from the received search query, acquires the expiration date corresponding to the hash value from the cache management table 204, that is, the expiration date corresponding to the search query, and the primary search result and Response cache expiration date is determined. If the search has been performed with the same search query as the received search query, the expiration date can be acquired. If the search query has not been searched with the same search query as the received search query, the expiration date cannot be acquired. The calculated hash value is stored in the cache management table 204 with an expiration date.

  The syntax analysis unit 103, the SQL generation unit 104, and the search execution unit 105 generate an SQL from the search query when the expiration date corresponding to the hash value does not exist or when the hash value has expired, and the RDBMS2 And execute the search. Specifically, the syntax analysis unit 103 creates a syntax tree from the search query, the SQL generation unit 104 generates SQL from the syntax tree, and executes a search on the RDBMS 2 using the SQL generated by the search execution unit 105. .

  The document acquisition unit 106 acquires fid and fver from the ResultSet object 205 that is a search result obtained from the RDBMS 2, and acquires a document (fragment) corresponding to fid from the fragment cache storage unit 111 or the RDBMS 2. The ResultSet object 205 includes fid, fver, and oid information indicating the storage location of the entity data. If a fragment of fid that matches the fid acquired from the ResultSet object 205 exists in the fragment cache storage unit 111, and the fver acquired from the ResultSet object 205 is less than or equal to the fver of the fragment cache storage unit 111, the fragment corresponding to fid Is obtained from the fragment cache storage unit 111. When the matching fragment of fid does not exist in the fragment cache storage unit 111, or when the fver acquired from the ResultSet object 205 is larger than the fver of the fragment cache storage unit 111, the entity data is obtained from the RDBMS2 according to the oid of the ResultSet object 205. get. The acquired entity data is stored in the fragment cache storage unit 111.

  The response data generation unit 107 generates, as response data, a document obtained by the document acquisition unit 106 delimited and combined with a delimiter. The generated response data is stored in the response cache storage unit 110 together with the hash value.

  The response data return unit 108 returns the response data to the terminal 3. When the hash value generated from the search query is within the validity period and the response data corresponding to the hash value exists in the response cache storage unit 110, the response data return unit 108 corresponds to the hash value from the response cache storage unit 110. The response cache (response data) to be read is read and returned to the terminal 3. If the hash value has expired or the response data corresponding to the hash value does not exist in the response cache storage unit 110, the response data generated by the response data generation unit 107 is returned to the terminal 3.

  The response cache storage unit 110 associates the hash value calculated from the search query with the response data obtained from the search query and stores it as a response cache.

  The fragment cache storage unit 111 stores a document (fragment) corresponding to fid as a fragment cache together with fid and fver.

  FIG. 2 is a schematic diagram illustrating the remence cache and the fragment cache. The response data shown in the figure is obtained by searching for information on programs A and B. There can be one or more fragments per program. In FIG. 2, program metadata of program A is fragments a1 to a3, and program metadata of program B is fragments b1 to b3. Fragments a1 to b3 are stored in the document table 202 in association with fid and fver. The response data is obtained by combining fragments a1 to b3, which are program metadata return units, obtained by searching for information on programs A and B.

  The response cache storage unit 110 caches response data in search request units. Specifically, the response cache storage unit 110 stores the hash value calculated from the search query for searching for information on the programs A and B and the response data obtained by combining the fragments a1 to b3 in association with each other.

  The fragment cache storage unit 111 caches fragments in units of fragments. Specifically, the fragment cache storage unit 111 stores the fragments a1 to b3 acquired from the RDBMS2 by the document acquisition unit 106 in association with fid and fver. fid is an identifier uniquely assigned to each of the fragments a1 to b3. fver is a value representing the version of the fragment. For example, when fragment a1 which is program metadata of program A is updated, fver is updated from 1 to 2. The update time may be simply used as fver.

  Next, the processing flow of the search device 1 will be described.

  FIG. 3 is a flowchart showing the flow of processing of the search device 1.

  When receiving the search query from the terminal 3, the query receiving unit 101 generates a hash value from the search query (step S101).

  Then, the cache expiration date determination unit 102 acquires the expiration date corresponding to the hash value from the cache management table 204 of the RDBMS 2 (step S102), and determines whether there is an expiration date corresponding to the hash value (step S103).

  When the expiration date corresponding to the hash value can be acquired, the cache expiration date determination unit 102 determines whether the expiration date is within the expiration date (step S104). If it has expired, the corresponding record is deleted from the primary search result table 203 and the cache management table 204 (step S105).

  If it is determined in step S103 that there is no expiration date corresponding to the hash value, or if the expiration date acquired in step S104 is determined to have expired, the syntax analysis unit 103, SQL generation is performed using the received search query. The primary search is performed on the RDBMS 2 by the unit 104 and the search execution unit 105 (step S106). The primary search is a search for the content ID corresponding to the search condition from the attribute table 201. Specifically, the content ID corresponding to the attribute in the attribute table 201 is searched according to the search condition. A list (cridlist) of content IDs corresponding to the search condition, which is the primary search result, is stored in the primary search result table 203 together with the hash value generated from the search query.

  On the other hand, when the expiration date corresponding to the hash value is within the time limit, it is determined whether or not response data corresponding to the hash value exists in the response cache storage unit 110 (step S107).

  When the response data corresponding to the hash value exists in the response cache storage unit 110, the response data return unit 108 acquires the response data from the response cache storage unit 110 (step S108) and returns it to the terminal 3 (step S108). S113).

  After performing the primary search or when there is no response data corresponding to the hash value, a secondary search is performed (step S109). The secondary search combines the primary search result of the primary search result table 203 and the document table 202 and acquires a document corresponding to a content ID list (cridlist) as a ResultSet object 205. Specifically, a table in which a list of content IDs (cridlist) obtained by the primary search and the document table 202 are internally joined by content ID (crid) is acquired.

  After the secondary search, the document acquisition unit 106 refers to the ResultSet object 205 obtained by the secondary search, and acquires a document from the fragment cache storage unit 111 or the RDBMS 2 (step S110). Details of step S110 for obtaining a document will be described later.

  Then, the response data generation unit 107 combines the acquired documents to generate response data (step S111), and stores the generated response data in the response cache storage unit 110 together with the hash value (step S112).

  Then, the response data generated by the response data return unit 108 is returned to the terminal 3 (step S113).

  Next, the flow of processing for acquiring a document will be described.

  FIG. 4 is a flowchart illustrating a flow of processing in which the document acquisition unit 106 acquires a document.

  First, the document acquisition unit 106 acquires one fid and fver from the ResultSet object 205 (step S201), and determines whether the fid and fver have been acquired (step S202). If fid and fver cannot be acquired, all the documents corresponding to the search condition have been acquired, and the process for acquiring the documents ends.

  Then, it is determined whether or not a fragment corresponding to the acquired fid exists in the fragment cache storage unit 111 (step S203). When the fragment corresponding to the acquired fid exists in the fragment cache storage unit 111, the acquired fver is compared with the fever of the fragment cache stored in the fragment cache storage unit 111 (step S204).

  If the acquired fver is less than or equal to the fver of the fragment cache, the document (fragment) is acquired from the fragment cache storage unit 111 (step S205).

  On the other hand, if the acquired fver is larger than the fver of the fragment cache, the document corresponding to the oid of the ResultSet object 205 is acquired from the RDBMS 2 (step S206). The acquired document is stored in the fragment cache storage unit 111 together with fid and fver (step S207).

  After the document is acquired, the process returns to step S201, and these processes are performed for all fids held in the ResultSet object 205.

  As described above, according to the present embodiment, individual documents constituting response data are stored in the fragment cache storage unit 111, and the document acquisition unit 106 stores the acquired document in the fragment cache storage unit 111. If it exists, the document is acquired from the fragment cache storage unit 111. If it does not exist, the document is acquired from the RDBMS2, thereby minimizing the number of times the document is acquired from the RDBMS2, and the search response time is reduced. It can be shortened.

  According to the present embodiment, the document version is stored in the fragment cache storage unit 111, and the document acquisition unit 106, when the document version in the fragment cache storage unit 111 is older than the document version in the RDBMS2, is RDBMS2 As a result, the cached document and the RDBMS2 document do not become inconsistent.

  In addition, by caching the response data for each search request in the response cache storage unit 110, the response time can be further shortened when the same search query is received.

DESCRIPTION OF SYMBOLS 1 ... Search apparatus 101 ... Query receiving part 102 ... Cache expiration date determination part 103 ... Syntax analysis part 104 ... SQL generation part 105 ... Search execution part 106 ... Document acquisition part 107 ... Response data generation part 108 ... Response data return part 110 ... Response cache storage unit 111 ... Fragment cache storage unit 2 ... RDBMS
201 ... Attribute table 202 ... Document table 203 ... Primary search result table 204 ... Cache management table 205 ... ResultSet object 3 ... Terminal

Claims (7)

A search device for searching for a document corresponding to a search condition from a database storing documents,
Receiving means for receiving a search query;
Search means for searching the database using the search query, and for obtaining a search result having storage location information indicating a storage location of the document and an identifier of the document with respect to a document corresponding to a search condition of the search query;
First cache storage means for storing a previously acquired document and an identifier of the document;
It is determined whether or not a document corresponding to the document identifier of the search result exists in the first cache storage unit. If the document exists in the first cache storage unit, the first cache storage unit If the document is acquired and does not exist, acquisition means for acquiring the document from the database based on the storage location information of the search result;
Storage means for storing the document acquired from the database together with an identifier of the document in the first cache storage means;
Generating means for combining the one or more documents acquired by the acquiring means to generate response data;
Return means for returning the response data;
A search device comprising: The search result has a version of the document;
The first cache storage means stores a version of the document,
If the version in the search result is the same as or older than the version in the first cache storage unit, the acquisition unit acquires the document from the first cache storage unit, and the version in the search result Is newer than the version in the first cache storage means, obtains the document from the database based on the storage location information of the search result,
The search device according to claim 1, wherein the storage means further stores the version of the document acquired from the database in the first cache storage means. An expiration date management table that holds an expiration date corresponding to the search query;
Second cache storage means for storing response data previously returned corresponding to the search query;
A determination unit that acquires the expiration date corresponding to the search query, can acquire the expiration date, and determines that the response data is acquired from the second cache storage unit when the expiration date is within the expiration date; ,
If the determination means determines that the response data is acquired from the second cache storage means, the return means acquires and returns the response data corresponding to the search query from the second cache storage means, 3. The search device according to claim 1, wherein when the response data generated by the generation unit is returned, the response data is stored in the second cache storage unit in association with the search query. A search method in which a search device searches the document corresponding to a search condition from a database storing documents,
Receiving a search query;
Searching the database using the search query, and obtaining a search result having storage location information indicating an identifier of the document and a storage location of the document with respect to a document corresponding to a search condition of the search query;
It is determined whether or not a document corresponding to the document identifier of the search result exists in the first cache storage unit that stores the previously acquired document and the identifier of the document, and the document is stored in the first cache storage unit. Obtaining the document from the first cache storage means if present, and obtaining the document from the database based on storage location information of the search result if not present;
Storing the document obtained from the database together with an identifier of the document in the first cache storage unit;
Combining one or more of the documents acquired by the acquisition means to generate response data;
Returning the response data;
A search method characterized by comprising: The search result has a version of the document;
The first cache storage means stores a version of the document,
The obtaining step obtains the document from the first cache storage means when the version in the search result is the same as or older than the version in the first cache storage means, If the version is newer than the version in the first cache storage means, the document is acquired from the database based on the storage location information of the search result;
5. The search method according to claim 4, wherein the storing step further stores the version of the document acquired from the database in the first cache storage unit. The expiration date corresponding to the search query is acquired from an expiration date management table that holds an expiration date corresponding to the search query, the expiration date can be acquired, and if the expiration date is within the expiration date, the search query Determining to obtain the response data from the second cache storage means for storing the response data previously returned in correspondence;
The returning step acquires and returns the response data corresponding to the search query from the second cache storage unit when it is determined in the determining step that the response data is acquired from the second cache storage unit. 6. The search according to claim 4, wherein when the response data generated in the generating step is returned, the response data is stored in the second cache storage unit in association with the search query. Method.   A search program for causing a computer to execute the search method according to claim 4.

Images