JP4550876B2 - Structured document retrieval system and program - Google Patents

Description

  The present invention relates to a structured document retrieval system for retrieving a structured document stored in a database, and more particularly to a structured document retrieval system and a program suitable for retrieval with a plurality of identification conditions.

  Many database management systems have a “same look search” function as part of the search function. The same view search is a search that specifies a search keyword and a character type or a character string that is identified with the search keyword. Examples of commonly used identification conditions are given below.

(1) Alphanumeric equivalence: Alphanumeric uppercase and lowercase letters (represented as case) and full-width (full-width characters) and half-width (half-width characters) (represented as width) are regarded as the same.

(2) Kana equation: Hiragana and Katakana (represented as kana), and hiragana and katakana uppercase and lowercase (represented as kanacase) are identified.

(3) Equivalent variant characters ・ ・ ・ Identify variant characters of kana and kanji.

(4) Identify synonyms ... Identify synonyms.

  For example, when the search keyword is “A”, if the alphanumeric identification condition is specified, “A”, “A”, and “A” will also be hit. If the kana identification condition is specified, “A A”, “A A”, and “A A” will also be hit. When multiple identification conditions are specified at the same time, more vocabulary hits are added.

  In this way, in the same-look search, in addition to the vocabulary that completely matches the search keyword, the vocabulary that matches the same-view condition is also hit. For this reason, the same-lookup search is convenient for searches that take into account fluctuations, but it also searches for vocabularies that have been developed according to the same-view conditions (so-called same-view development), which increases the search processing cost. . For example, Patent Documents 1 and 2 propose a method for solving this problem by devising index information. Patent Document 1 proposes a method of suppressing processing costs by dividing index information to be used for searches that are not identified and searches that are not identified. Patent Document 2 proposes a method of reducing processing costs by devising a hash value and managing identifiable vocabulary as the same index information.

  In recent years, documents having a logical structure (that is, structured documents) have been managed by database management systems. In the structured document, the logical structure of the document is indicated by a tag described in the document. A structured document in which a logical structure is expressed using this tag is suitable for processing by a computer.

  As a means for describing data using tags, XML (Extensible Markup Language) is widely used. XML has features such as hierarchization of data by means of tagged tags and free expandability of structure. A database called XML database (XMLDB) is known as an XML utilization technology that makes use of these features. The XML database is controlled by a database management system called an XML database management system (XMLDBMS), and provides functions for storing XML documents and retrieving XML documents (structures specified therein). Even in such a database management system, it is possible to search for an XML document (structured document) that satisfies the same view condition by applying the same view search.

A standardized query language is often used to search for XML documents (elements therein). The main query languages are XPath and XQuery. XPath is used to perform a search by specifying the position of an element (node) in an XML document.
JP 2004-199282 A JP 2006-106896 A

  In the same-sight search, there may be a case where it is necessary to specify a condition for switching whether or not to identify the same for each same-sight condition, but conversely, a condition specification for specifying whether or not to identify all the same-view conditions together is necessary. In some cases. An example is a web search service.

  In the prior art, it is necessary to individually specify a plurality of identical viewing conditions even in the latter case. Therefore, it takes time to write a query (inquiry), and there is a problem that the query becomes long and it is difficult to interpret the meaning. It is also necessary to know all the query grammars for specifying identification conditions.

  A specific example is shown using XQuery. Here, for convenience of explanation, it is assumed that there are only two identification conditions: “alphanumeric identification” and “kana identification”, and alphanumeric identification is “case / width (in) sensitive”. Is described as “kana / kanacase (in) sensitive”.

When specifying to identify all the same identification conditions, the query (search expression) is, for example,
/ item / name [./text () ftcontains “Employee A” case insensitive width insensitive kana insensitive kanacase insensitive]
become that way. Here, “ftcontains” means an XQuery full-text search operator. In addition, “case insensitive”, “width insensitive”, “kana insensitive”, and “kanacase insensitive” are the parts that specify the same identification condition, and these four identification condition specification parts make the query longer. Yes.

  On the other hand, in the same-look search, regardless of whether or not the hit is actually made, the search keyword is looked-up according to the designation of the same-view condition. For this reason, the more the same viewing conditions that are specified at the same time, the more the number of expansions, and the search processing cost increases. For example, in the case of a two-character search keyword “A” that includes alphanumeric characters and hiragana, the number of expansions is four if the single designation of alphanumeric identification or kana identification is single. However, the number of expansions is increased to 16 if the alphanumeric identification and kana identification are specified simultaneously.

  Patent Documents 1 and 2 propose methods for reducing processing costs. However, in the methods proposed by Patent Documents 1 and 2, the search keywords are identified in accordance with the identification conditions. For this reason, when the expanded vocabulary is not hit, the processing cost required for the expansion is wasted.

  The present invention has been made in consideration of the above circumstances, and the purpose of the present invention is to perform an identification search according to a plurality of identification conditions based on the search expression even if the search expression does not describe a plurality of identification conditions. It is an object of the present invention to provide a structured document search system and program that can be performed.

  According to one aspect of the present invention, a structured document search system including a database that stores a plurality of structured documents is provided. The structured document search system analyzes the identification formula information storage means for storing a plurality of identification conditions for identifying the search keyword and a search expression included in the search request given from the client terminal, Search processing means for performing a search process on the structured document stored in the database based on the search formula; and batch identification condition for the search formula to collectively specify all identification conditions , The same-view expansion means for expanding the search keyword included in the search formula based on all of the plurality of the same-view conditions stored in the same-view condition information storage means. The search processing unit searches a structured document including a structure that matches a search condition indicated by the search expression from the database based on the search keyword included in the search expression and the expanded search keyword. .

  According to the present invention, whether or not all of the plurality of identification conditions stored in the identification condition information storage unit are included in the retrieval expression only by including (description) the collective identification condition in the retrieval expression. As described above, the same-view search according to the plurality of same-view conditions can be performed based on the search formula. For this reason, it is easy for the user to describe the search formula, and since the search formula is shortened, the meaning can be easily interpreted. Further, even if the user does not know each of the plurality of identification conditions automatically specified by the batch identification condition, the user can search the structured document search system using all of the plurality of identification conditions. It can be requested from the terminal. Further, even if the number of identical conditions automatically specified by the collective identical condition is increased, that is, even when a new identical condition is added to the identical condition information storage unit, there is no need to change the search formula. .

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a block diagram showing a hardware configuration of a client-server system including a structured document search system 50 according to an embodiment of the present invention. The client-server system mainly includes a database server 10 and a plurality of client terminals (clients). The plurality of client terminals include a client terminal 20. On the client terminal 20, an application (application program) that uses the database server 10 operates. A plurality of client terminals including the client terminal 20 are connected to the database server 10 via a network 30 such as a local area network (LAN).

  In FIG. 1, client terminals other than the client terminal 20 are omitted. Therefore, in the following description, it is assumed that only the client terminal 20 uses the structured document search system 50. However, other client terminals can use the structured document search system 50 in the same manner as the client terminal 20.

  The database server 10 is a computer (database server computer) having a memory 11 such as a main memory. The database server 10 is connected to an external storage device 40 such as a hard disk drive. The external storage device 40 stores a database management program 41 and a database 42. In the present embodiment, the structured document search system 50 is realized by the database server 10 and the external storage device 40 (the database 42 stored therein).

  The database management program 41 is used for management of the database 42 by the database server 10 and search processing based on an inquiry from the client terminal 20 using, for example, XQuery. The database 42 is an XML document database (structured document database) that stores, for example, XML documents (electronic documents in XML format) as structured documents.

  FIG. 2 is a block diagram mainly showing a functional configuration of the structured document search system 50 shown in FIG. The structured document search system 50 includes a database management system 51 and a database 42. In addition to the set of XML documents, the database 42 stores an index used to search for XML documents stored in the database 42.

The database management system 51 includes a request processing unit 52, an identification condition management unit 53, a search processing unit 54, an index management unit 55, a document registration unit 56, and a database operation unit 57.
The request processing unit 52 receives a request (command) from the client terminal 20, determines the type of the request, and determines the request based on the determination result, the request processing unit 52, the identification condition management unit 53, and the search processing unit. 54, and functions as an input interface to be sent to the index management unit 55 or the document registration unit 56. Here, if the request from the client terminal 20 is the same viewing condition setting request (details will be described in the second modification), the request processing unit 52 sends the same viewing condition setting request to the same viewing condition management unit 53. ,
If a search request, the search request is sent to the search processing unit 54. The request processing unit 52 sends the index setting request to the index management unit 55 if the request from the client terminal 20 is an index setting request, and sends the document registration request to the document registration unit 56 if the request is a document registration request. The request processing unit 52 also functions as an output interface that returns a response from the identification condition management unit 53, the search processing unit 54, the index management unit 55, or the document registration unit 56 to the request from the client terminal 20. To do.

  The same-view condition management unit 53 manages the same-view condition, for example, the same-view condition for each session used for communication between the structured document search system 50 (internal database management system 51) and the client terminal 20. The identical viewing condition management unit 53 includes an identical viewing condition storage unit 530. The same viewing condition information storage unit 530 stores the same viewing condition information 531 used for managing application / non-application (that is, whether or not to view) each of the same viewing conditions for each session described above. The same viewing condition information storage unit 530 is assumed to be realized using a partial storage area in the memory 11 of the database server 10. The identical viewing condition information storage unit 530 may be realized using a partial storage area of the external storage device 40.

  FIG. 3 shows an example of the same viewing condition information 531. In the following, for convenience of explanation, it is assumed that there are only two identification conditions, “alphanumeric identification (case / width)” and “kana identification (kana / kanacase)”.

  In this embodiment, the same viewing condition information 531 is managed in a table format as shown in FIG. The same viewing condition information 531 includes a session number and each same viewing condition as a column element (column item). For each session (in this embodiment, a connected session, that is, an established session) used for communication with a client terminal, the same-view condition information 531 includes the session number of the session and the same-view of each same-view condition. Whether or not (applied / non-applied) is set as a value (table element value). Information consisting of a set of a session number and a value indicating whether or not each of the same viewing conditions is identified for each session is referred to as session-corresponding identification information.

  When the session is not established, the same viewing condition information unique to the session (session-corresponding viewing condition information) is empty. When the session is established and the request processing unit 52 notifies the same viewing condition management unit 53 to that effect, information regarding the session is added to the same viewing condition information 531 by the same viewing condition management unit 53. Here, each identification condition in identification information specific to the established session is set to “Yes” (identify). In the example of FIG. 3, depending on the session (session number), “do not” (not identified), which means the opposite setting of “yes”, is also described. However, in this embodiment, when the session is established All are set to “Yes”. When the session is disconnected and the fact processing unit 52 notifies the same viewing condition management unit 53 to that effect, the information regarding the session is deleted by the same viewing condition management unit 53.

  Referring to FIG. 2 again, the search processing unit 54 receives a search request (query) sent from the client terminal 20 via the request processing unit 52, and performs a search process according to the search formula included in the query.

  The index management unit 55 manages an index used for search processing by the search processing unit 54. The index management unit 55 creates an index associated with a node (element) included in the XML document and registers it in the database 42 when the document registration unit 56 registers the XML document.

  The document registration unit 56 receives a document registration request from the client terminal 20 via the request processing unit 52, and performs document registration processing for registering (storing) the XML document specified in the document registration request in the database 42. The database operation unit 57 functions as an interface that enables the search processing unit 54, the index management unit 55, and the document registration unit 56 to access the database (database file) 42, and performs processing on the database 42.

  In this embodiment, the request processing unit 52, the same-sight condition management unit 53, the search processing unit 54, the index management unit 55, the document registration unit 56, and the database operation unit 57 are connected to the external storage device 40 by the database server 10 in FIG. It is assumed that the stored database management program 41 is implemented by reading it into the memory 11 in the server 10 and executing it. The program 41 can be stored in advance in a computer-readable storage medium such as a compact disk or ROM and can be distributed. Further, this program 41 may be downloaded to the database server 10 via the network 30.

Next, the operation of the structured document search system 50 will be described.
<Document registration process>
First, document registration processing in the structured document search system 50 will be described with reference to the flowchart of FIG.

  Now, it is assumed that the user designates an XML document to be registered in the database 42 and performs an operation on the client terminal 20 to instruct the registration of the XML document. Then, the client terminal 20 sends a registration request (document registration request) for registering the designated XML document in the database 42 to the structured document search system 50 via the network 30 (step S1).

  Upon receiving the registration request from the client terminal 20, the request processing unit 52 passes the registration request to the document registration unit 56 and requests registration of an XML document (XML document) (step S2). When the document registration unit 56 receives a registration request from the client terminal 20 via the request processing unit 52, the document registration unit 56 starts parsing (syntax analysis) of the XML document designated by the request (step S3).

Each time an element (node) is extracted from the XML document by parsing the XML document, the document registration unit 56 performs the following process on the node (step S4).
First, the document registration unit 56 passes the extracted node information, for example, the information (node information) including the extracted node structure (representing a path) to the index management unit 55 to perform index registration processing. When the content (value) of the extracted node (element) is text, a node corresponding to the content of the node, that is, a node immediately below the node is called a text node. As described above, when the content of the extracted node (element) is text, the node information includes the structure (path) of the node and the value of the text node immediately below the node.

  Upon receiving the node information from the document registration unit 56, the index management unit 55 creates an index including the structure (path) in the node information, and performs an index registration process for additionally registering the index in the database 42 (step). S5). In the present embodiment, when the node information includes a text node value, the created index includes a set of a structure (path) and a text node value.

  Next, the document registration unit 56 performs document storage processing for storing the extracted node (that is, a part of the XML document) in the database 42 (step S6). Within the database 42, a plurality of XML documents stored in the database 42 is a partial document that forms part of the structure of one virtual XML document, and the root node (/) of the virtual XML document. It is positioned below

  When the document registration process (step S6) is performed, the document registration unit 56 has completed the parsing of the XML document requested from the client terminal 20, that is, has completed the process for all the nodes included in the XML document. (Step S7). If an unprocessed node remains (step S7), the document registration unit 56 returns to step S4 and continues the process for the next node. On the other hand, if there are no unprocessed nodes left, that is, if all the requested XML documents have been parsed (step S7), the search processing unit 54 ends the document registration process.

  The order of nodes extracted from the XML document matches the order of nodes appearing in the XML document. Therefore, in the document registration process described above, nodes are processed in the order in which they appear in the XML document.

Here, the XML document requested to be registered by the client terminal 20 is
<item>
<id> 0001 </ id>
<name> Employee A </ name>
</ item>
Suppose that

  FIG. 5 shows an example of an index created when the XML document is registered in the database 42. In the example of FIG. 5, the index includes a set of a node structure (path) extracted from the registered XML document and a text node value immediately below it (on the tree structure of the XML document).

  For example, the index for the <id> node is the <id> node (element) structure “/ item / id” and the text node value “0001” that is the content of the <id> node, that is, the <id> node The value “0001” of the text node immediately below is included. Similarly, the index related to the <name> node includes the structure “/ item / name” of the <name> node and the value “Mr. employee A” of the text node immediately below the <name> node. However, for a node that does not have a text node immediately below it, such as an <item> node, the symbol “-” meaning an empty character string is used instead of the value of the text node.

<Search process>
Next, search processing in the structured document search system 50 will be described with reference to the flowchart of FIG.

  Now, when a user operates the client terminal 20, a search request is sent from the client terminal 20 to the structured document search system 50 using the currently established session via the network 30. (Step S11). This search request includes a search expression.

  Upon receiving the search request from the client terminal 20, the request processing unit 52 passes the search request to the search processing unit 54 and requests a search (step S12). The search processing unit 54 analyzes the search formula included in the search request (step S13), so that the search formula includes “batch equating condition”, that is, the search formula includes “batch equate”. It is determined whether or not “condition” is designated (step S14). The collective identification condition means a search condition for collectively specifying all of a plurality of preset identification conditions. In the present embodiment, it is assumed that the description of the character string “all insensitive” is used for the collective identification condition.

  If the collective identification condition “all insensitive” is specified in the search formula (step S14), the search processing unit 54 uses the session number of the session used to transmit the search request including the search formula. The viewing condition management unit 53 is inquired about the same viewing condition information (session-matching same viewing condition information) of the session (step S15). Then, the same viewing condition management unit 53 selects only the same viewing condition information unique to the session having the session number inquired from the search processing unit 54 from the same viewing condition information 531 stored in the same viewing condition information storage unit 530. Is extracted and returned to the search processing unit 54. Thereby, in step S15, the search processing unit 54 sets the same viewing condition information (session-corresponding viewing condition information) unique to the target session, that is, a plurality of sets that are all identified with the target session. Information indicating the same viewing condition is acquired. Note that the search processing unit 54 may directly acquire the same viewing condition information 531 stored in the same viewing condition information storage unit 530 to acquire the same viewing condition information unique to the target session. .

  Next, the search processing unit 54 sets each identical viewing condition as a retrieval condition indicated by the analyzed retrieval formula in accordance with the acquired identical viewing condition information (step S16). The setting of the search condition according to the identification condition information will be described with a specific example assuming that the search expression (query) is an XQuery expression.

First, in the session with session number 1 (hereinafter referred to as session # 1), the following search expression (query)
/ item / name [./text () ftcontains “Employee A” all insensitive]
It is assumed that a search request including is executed. At this time, it is assumed that the same viewing condition information 531 stored in the same viewing condition information storage unit 530 is in the state of FIG.

  In the above query, the batch identification condition “all insensitive” is specified. Therefore, the search processing unit 54 obtains the same viewing condition information (session-corresponding viewing condition information) unique to the session # 1 by making an inquiry to the same viewing condition management unit 53 (step S15).

  As is clear from FIG. 3, the identification condition information unique to session # 1 is to “equalize” (identify) both alphanumeric identification and kana identification (“kana insensitive” and “kanacase insensitive”). Includes set identification conditions (alphanumeric identification conditions and pseudonym identification conditions). In this case, the search processing unit 54 recognizes the collective identification condition “all insensitive” in the query as alphanumeric characters (“case insensitive” and “width insensitive”) and also identifies the kana (“kana insensitive”). And “kanacase insensitive”). The search processing unit 54 automatically replaces “all insensitive” in the query with “case insensitive width insensitive kana insensitive kanacase insensitive”.

As a result, the above query becomes
/ item / name [./text () ftcontains “Employee A” case insensitive width insensitive kana insensitive kanacase insensitive]
Is equivalent to

  As described above, the search processing unit 54 replaces the batch identification condition “all insensitive” according to the acquired identification condition information, and indicates that the alphanumeric identification and the pseudonym identification are both performed. Conditions (“case insensitive” and “width insensitive”) and pseudonym identification conditions (“kana insensitive” and “kanacase insensitive”) are set (step S16). In the present embodiment, when both the batch identification condition and the individual identification condition are specified by the query, the individual identification condition is ignored (treated as an error).

  When step S16 is executed by the search processing unit 54, the index management unit 55 functions as an identical development means, and the search keyword included in the query (search expression) analyzed by the search processing unit 54 and the search keyword The search keyword is developed in accordance with the individual identification conditions set in step S16 (step S17). Next, the index management unit 55 functions as an index search processing unit, and uses, as a search result, a list of nodes that match the search condition using the index stored in the database 42 for all search keywords after expansion. Obtain (step S18). Here, since “/ item / name” is designated as a search target path by the above query (search formula), an index including the path “/ item / name” is used for the search in step S18. The search processing unit 54 returns the search result using the index by the index management unit 55 to the client terminal 20 by the request processing unit 52 (step S19).

  The index stored in the database 42 is copied to the memory 11 in the database server 10 when the structured document search system 50 is activated, and this copied index is used for the search processing by the index management unit 55. As a result, it is possible to speed up the search process. In addition, the search processing unit 54 can have one or both of the function as the same-view developing unit and the function as the index search processing unit of the index management unit 55. In addition, it is possible to provide a function independent from the search processing unit 54 as a function of developing the same view.

  On the other hand, if the batch identification condition is not specified in the search formula included in the search request (step S14), steps S15 and S16 are skipped and step S17 is executed. In this case, the search keyword is expanded according to the same viewing condition directly specified by the query included in the search request from the client terminal 20.

  According to this embodiment, since it is not necessary to describe individual identification conditions in the search formula, the description of the search formula is easy for the user, and the search formula is shortened, so that the meaning can be easily interpreted. Further, even if the user does not know about the individual identification conditions, the structured document search system 50 can be made to perform a search using all of the preset identification conditions simply by using the collective identification conditions. it can. Further, according to the present embodiment, even if the number of identification conditions that can be collectively specified increases, if the new conditions are the same as or not identical to the already set identification conditions, the search formula There is no need to change.

[First Modification]
In the above-described embodiment, by setting (description) the collective identification condition in the query, it is possible to collectively specify all of the plurality of preset identification conditions. However, when the same viewing condition unnecessary for the search is included in a plurality of the same viewing conditions that are collectively specified, the batch same viewing condition cannot be used.

  Therefore, a first modification example of the above-described embodiment that enables designation of the same viewing condition to be excluded from the batch simultaneous viewing condition will be described. In this first modification, the query (search formula) includes the same identification condition to be excluded from the collective identification condition, that is, the same identification condition “except XXX” (“XXX” is excluded from the collective identification designation target). It is possible to describe a character string indicating the same viewing condition such as “case” and “width”. The same viewing condition that is not subject to the collective viewing condition is referred to as a same viewing exclusion condition.

<Search process>
Hereinafter, the operation of the first modification of the above embodiment will be described with reference to the flowchart of FIG. 7 focusing on the differences from the above embodiment, taking the search processing in the structured document search system 50 as an example. In FIG. 7, the same steps as those in the flowchart of FIG. 6 are denoted by the same reference numerals.

  First, it is assumed that when the user operates the client terminal 20, a search request is transmitted from the client terminal 20 to the structured document search system 50 via the network 30 (step S11). The search formula included in this search request is assumed to include the collective identification condition as in the above embodiment. In this case, steps S12 to S16 are performed in the same manner as in the above-described embodiment, and individual identification conditions for which the same identification is designated are set.

  Then, the search processing unit 54 determines whether or not the same eye exclusion condition is specified in the query (search formula) analyzed in step S13 (step S21). In the first modification, it is assumed that this query specifies (includes) the same-identification exclusion condition in addition to the collective identification condition.

  When the equated exclusion condition is specified in the analyzed query (step S21), the search processing unit 54 has a plurality of equated conditions set in step S16, that is, a plurality of equated conditions specified in a lump. Only the same vision condition designated by the same vision exclusion condition is excluded (step S22). A process for excluding the same vision condition specified by the same vision exclusion condition will be described with a specific example.

First, in session # 1, the following search expression (query)
/ item / name [./text () ftcontains “Employee A” all insensitive except case]
It is assumed that a search request including is executed. At this time, it is assumed that the same viewing condition information 531 stored in the same viewing condition information storage unit 530 is in the state of FIG.

  In the above query, the batch identification condition “all insensitive” is specified. Therefore, as in the above embodiment, “all insensitive” is automatically replaced with “case insensitive width insensitive kana insensitive kanacase insensitive” (step S16).

  Further, in the above query, the sameness exclusion condition “except case” is specified (step S21). Therefore, the search processing unit 54 excludes alphanumeric case insensitivity (case insensitive) from “case insensitive width insensitive kana insensitive kanacase insensitive” (step S22). That is, the search processing unit 54 automatically replaces “case insensitive width insensitive kana insensitive kanacase insensitive” with “width insensitive kana insensitive kanacase insensitive” from which alphanumeric case insensitivity (case insensitive) is excluded.

As a result, the above query becomes the following query:
/ item / name [./text () ftcontains “Employee A” width insensitive kana insensitive kanacase insensitive]
Is equivalent to

  Subsequent operations are the same as those in the above embodiment, and the index management unit 55 and the search keyword included in the query (search expression) analyzed by the search processing unit 54 and the individual identification conditions currently set ( Here, the search keyword is expanded in accordance with a plurality of the same viewing conditions specified in batch and the remaining viewing conditions from which the same viewing conditions specified in the same viewing exclusion conditions are excluded (step S17). Note that if the same-eye exclusion condition is not designated (step S21), step S22 is skipped and step S17 is executed. When step S17 is executed, steps S18 and S19 are executed as in the above embodiment.

  According to the first modification of the above embodiment, it is possible to set the same-view condition that is not subject to the batch identification specification by the same-view exclusion condition. But you can use batch identification. For this reason, the batch identification designation can be used more effectively. In addition, since the same-look exclusion condition can be set (described) in the query, the same-view condition can be changed in units of queries while using batch identification.

[Second Modification]
In the first modification of the above-described embodiment, when a search for excluding a desired identical viewing condition from a batch identification designation is necessary, the user needs to designate the identical exclusion condition by a retrieval formula each time. .

  Therefore, a description will be given of a second modification of the above-described embodiment that makes it possible to exclude a desired equating condition from being designated as a collective equating designation without designating the equating exclusion condition with a search expression. A feature of the second modified example is a setting of whether or not to identify (apply / not apply) for each of a plurality of identification conditions that are targets of batch identification specification associated with a session requested by the user. Is to be able to update.

<Identification condition setting process>
Hereinafter, the operation of the second modified example of the above embodiment will be described with reference to the flowchart of FIG.
First, when the user operates the client terminal 20, the client terminal 20 sends an identification condition setting request to the structured document search system 50 through the network 30 using the currently established session. (Step S31). This request for setting the same viewing condition includes information regarding whether or not to perform the same viewing for each predetermined viewing condition (applied / not applied).

  When receiving the same viewing condition setting request from the client terminal 20, the request processing unit 52 passes the same viewing condition setting request to the same viewing condition management unit 53 and requests the same viewing condition setting (step S32).

  Then, the same-view condition management unit 53 stores the same-view condition information stored in the same-view condition information storage unit 530 according to the information regarding whether or not to view (apply / not apply) included in the request for setting the same-view condition. 531, information on whether or not to identify (apply / not apply) for each of a plurality of identification conditions that are targets of batch identification specification associated with the corresponding session (session-corresponding identification information). Update (step S33). Here, the corresponding session refers to a session used for transmission of the same-view condition setting request from the client terminal 20.

  If it is specified in the identification condition setting request that only alphanumeric uppercase and lowercase identification conditions are not identified (not applicable), the execution of step S33 causes the alphanumeric identification conditions and the pseudonym identification conditions. Of these, only alphanumeric upper and lower case identification conditions are changed to not identify. This is equivalent to designating the identification exclusion condition “except case” each time by a query.

  According to the second modified example of the above-described embodiment, the plurality of identification conditions that are targets of batch identification specification associated with the session requested by the user are / are not identified (applied / non-applicable). Because the setting of (apply) can be updated, it is possible to continuously make the setting of the same viewing condition that is the target of the batch identification specification in the session. Therefore, the user does not need to set the sameness exclusion condition for each query.

[Third Modification]
In the embodiment described above (and the first and second modifications of the embodiment), the search keywords are unconditionally expanded according to the individual identification conditions set as the target of batch identification. However, if the expanded vocabulary is not hit, the processing cost required for the expansion is wasted.

  Therefore, a third modification of the above-described embodiment that can prevent the development of such meaningless search keywords will be described.

  FIG. 9 is a block diagram showing a configuration of the index management unit 55 applied in the third modification of the above embodiment. The index management unit 55 includes a character type information storage unit 550. The character type information storage unit 550 stores the character type (text character type) and the number of characters of the text node immediately below each hierarchical level of the XML document registered in the database 42, that is, for each structure (path) of the XML document. Character type information 551 used for management is stored.

  FIG. 10 shows an example of character type information 551. In the third modification, the character type information 551 is managed in a table format. The character type information 551 includes a path and a character type as column elements. Here, it is assumed that there are three types of characters: “alphanumeric characters”, “kana”, and “others”. The character type information 551 includes, for each path, a character string (XPath notation character string) representing the path and the number of characters for each character type of the text node immediately below the path of the XML document registered in the database 42 (value ( Table element value). Information consisting of a set of a character string representing the path and the number of characters for each character type for each path is referred to as path-corresponding character type information. The character type information 551 in FIG. 10 shows an example in which only the XML document mentioned in the above embodiment is registered in the database 42 for simplification. In the example of this XML document, the value of the text node immediately below the <id> node is “0001”, and the value of the text node immediately below the <name> node is “Employee A”. Therefore, in FIG. 10, the character type “alphanumeric, kana, other” associated with the path “/ item / id” is “4, 0, 0” and corresponds to the path “/ item / name”. The number of characters of the attached character type “alphanumeric characters, kana, etc.” is “1, 4, 0”.

  When no XML document is registered in the database 42, the character type information 551 is empty. Each time an XML document is registered in the database 42 and a new path (XPath notation) increases, a line of character type information 551 is added, and for each character type of a text node immediately below the path included in the XML document. A value (number of characters) is registered. In addition, for a path already registered in the character type information 551, the number of characters of the character type already registered is the same as the number of characters of the character type of the text node immediately below the path included in the newly registered XML document. Will be increased. Similarly, when an XML document is deleted from the database 42, the number of characters of the registered character type is decreased by the number of characters of the character type of the text node immediately below the path included in the deleted XML document. The In addition, for a path that no longer exists due to the deletion of the XML document, the line related to the path is deleted from the character type information 551.

<Document registration process>
Next, the operation of the third modified example of the above embodiment will be described with reference to the flowchart of FIG. 11 taking the document registration process in the structured document search system 50 as an example. In FIG. 11, the same steps as those in the flowchart of FIG. 4 are denoted by the same reference numerals.

  Now, it is assumed that the user designates an XML document to be registered in the database 42 and performs an operation on the client terminal 20 to instruct the registration of the XML document. Then, the client terminal 20 sends a registration request (document registration request) for registering the designated XML document in the database 42 to the structured document search system 50 via the network 30 (step S1).

  Upon receiving the registration request from the client terminal 20, the request processing unit 52 passes the registration request to the document registration unit 56 and requests registration of an XML document (XML document) (step S2). When the document registration unit 56 receives a registration request from the client terminal 20 via the request processing unit 52, the document registration unit 56 starts parsing the XML document specified by the request (step S3). Then, as in the above embodiment, each time an element (node) is extracted from the XML document (step S4), an index registration process by the index management unit 55 is performed for that node (step S5).

  Next, the index management unit 55 functions as a character type information management unit, and when the contents (values) of the nodes (elements) subjected to index registration are text, the character type is identified for the text and the number of characters is calculated. In the character type information 551 stored in the character type information storage unit 550, the number of characters of the identified character type associated with the path of the node is updated (for example, increased by the calculated number of characters) ( Step S40). 7 is different from the flowchart of FIG. 4 in that step S40 is added. It should be noted that the function as the character type information management means of the index management section 55 can be provided in a means independent from the index management section 55.

  When the character type information 551 is updated by the index management unit 55 (step S40), the document registration unit 56 performs document storage processing for storing the extracted node in the database 42 (step S6). If there is an unprocessed node remaining (step S7), the document registration unit 56 returns to step S4 and continues the process for the next node.

<Search process>
Next, search processing in the structured document search system 50 applied in the third modification will be described with reference to the flowchart of FIG. 12 with a focus on differences from the above embodiment. In FIG. 12, the same steps as those in the flowchart of FIG. 6 are denoted by the same reference numerals.

  First, it is assumed that when the user operates the client terminal 20, a search request is transmitted from the client terminal 20 to the structured document search system 50 via the network 30 (step S11). The search formula included in this search request is assumed to include the collective identification condition as in the above embodiment. In this case, steps S12 to S16 are performed in the same manner as in the above-described embodiment, and individual identification conditions for which the same identification is designated are set.

  Then, the index management unit 55 acquires character type information (path-corresponding character type information) specific to the search target path from the character type information 551 stored in the character type information storage unit 550 (step S50). In this step S50, the index management unit 55 functions as an identification condition excluding unit, and is based on the number of characters (number of characters) included in the acquired character type information, so that there is no wasted identification. A condition is specified, and the specified identical viewing condition is excluded from a plurality of identical viewing conditions designated collectively as identical. The flowchart of FIG. 12 differs from the flowchart of FIG. 6 in that step S50 is added. The detailed procedure of step S50 will be described later. The search processing unit 54 may perform the above step S50 by inquiring the index management unit 55 about character type information (path-corresponding character type information) specific to the search target path.

  On the other hand, if the batch identification condition is not specified in the search expression included in the search request (step S14), step S15 and step S16 are skipped, and step S50 is executed. Here, the above-identified identification conditions are excluded from the identification conditions directly specified by the search expression.

  When step S50 is executed, the index management unit 55 expands the search keyword according to the search keyword included in the analyzed query (search expression) and the identification conditions not excluded in step S50. (Step S17). Next, the index management unit 55 acquires, as a search result, a list of nodes that match the search condition using the index stored in the database 42 for all search keywords after expansion (step S18). Here, useless identification conditions that do not hit are excluded, and development of search keywords according to the useless identification conditions is suppressed. In other words, useless expansion of search keywords is suppressed. This speeds up the search.

<Process to exclude useless identification conditions>
Next, a detailed procedure of the processing in step S50 (that is, processing for removing useless identification conditions) by the index management unit 55 will be described with reference to the flowchart of FIG.

  First, the index management unit 55 acquires character type information (path-corresponding character type information) specific to the search target path from the character type information 551 stored in the character type information storage unit 550 (step S61). Next, the index management unit 55 determines whether the number (number of characters) of the character type “alphanumeric characters” in the acquired character type information is 0 (step S62). If the number of alphanumeric characters is 0 (step S62), the index management unit 55 assumes that no alphanumeric hit identification condition will be hit even if the alphanumeric identification condition is specified. (Step S63). That is, the index management unit 55 sets the alphanumeric identification condition as “not to be identified” (not applied). On the other hand, when the number of alphanumeric characters is not 0 (step S62), the index management unit 55 skips step S63.

  Similarly, the index management unit 55 determines whether or not the number (character number) of the character type “kana” in the acquired character type information is 0 (step S64). If the number of kana is 0 (step S64), the index management unit 55 excludes the kana equating condition from the equating condition, assuming that the kana equating condition does not hit even if the kana equating condition is specified. (Step S65). That is, the index management unit 55 sets the pseudonym identification condition as “not to be identified” (not applied). On the other hand, if the number of kana is not 0 (step S64), the index management unit 55 skips step S65. In addition, the process (steps S61 to S65) of removing the above-mentioned useless identification condition is performed by the search processing unit 54 inquiring the index management unit 55 for character type information (path-corresponding character type information) specific to the search target path. The search processing unit 54 can also perform this.

  According to the third modification of the above embodiment, useless expansion of search keywords is suppressed by eliminating useless identification conditions. Thereby, the search processing cost is reduced, and the search can be speeded up. In addition, since the process of eliminating useless identification conditions is automatically performed, the user can set a search expression for the identification search without considering the character types constituting the search keyword.

  In the above-described embodiment and its modification, it is assumed that the structured document is an XML document. However, the present invention can be similarly applied to structured documents other than XML documents such as SGML (Standard Generalized Markup Language) documents.

  In addition, this invention is not limited to the said embodiment or its modification example as it is, A component can be deform | transformed and embodied in the range which does not deviate from the summary in an implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment or its modification. For example, you may delete a some component from all the components shown by embodiment or its modification.

1 is a block diagram showing a hardware configuration of a client-server system including a structured document search system according to an embodiment of the present invention. FIG. 2 is a block diagram mainly showing a functional configuration of the structured document search system shown in FIG. 1. The figure which shows an example of the same viewing condition information applied in the embodiment. 6 is a flowchart showing a procedure of document registration processing in the embodiment. The figure which shows the specific example of the index applied in the embodiment. 6 is a flowchart showing a procedure of search processing in the embodiment. The flowchart which shows the procedure of the search process in the 1st modification of the embodiment. The flowchart which shows the procedure of the same viewing condition setting process in the 2nd modification of the embodiment. The block diagram which shows the structure of the index management part applied in the 3rd modification of the embodiment. The figure which shows an example of the character type information applied in the 3rd modification. 14 is a flowchart showing a procedure for document registration processing in the third modification. The flowchart which shows the procedure of the search process in the 3rd modification. The flowchart which shows the procedure of the process which excludes the useless identification condition in a 3rd modification.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Database server, 11 ... Memory, 20 ... Client terminal, 30 ... Network, 40 ... External storage device, 41 ... Database management program, 42 ... Database, 50 ... Structured document search system, 51 ... Database management system, 52 ... Request processing unit 53 ... Equal viewing condition management unit 54 ... Search processing unit 55 ... Index management unit 56 ... Document registration unit 530 ... Equal viewing condition information storage unit 531 ... Equal viewing condition information 550 ... Character type information Storage unit, 551... Character type information.

Claims (5)

A database for storing multiple structured documents;
Identification condition information storage means for storing a plurality of identification conditions for identification with the search keyword;
When the retrieval formula included in the retrieval request given from the client terminal includes a batch identification condition for collectively specifying all identification conditions, the plurality of items stored in the identification condition information storage unit equated expansion means for expanding a search keyword included in the search expression based on all identified conditions,
Search processing means for searching a structured document including a structure that matches a search condition indicated by the search formula from the database, based on the search keyword included in the search formula and the expanded search keyword;
When the identification viewing means expands a search keyword included in the search formula based on each of the identification conditions, the structured data stored in the database of character types unique to the identification conditions. Based on the presence / absence of characters corresponding to the structure indicated by the search expression among the characters included in the document, the presence / absence of the possibility of hitting is determined by the search under each of the same identification conditions. And the same condition exclusion means for excluding conditions ,
The structured document search system, wherein the same view development means excludes the same view conditions excluded by the same view condition exclusion means from the search keyword expansion target . For each of a plurality of character types predetermined character type information storage means to store information indicating whether a character of the character type included in the structured document stored in the database, for each structure of the structured document further comprising,
The equivalence condition exclusion means associates the presence or absence of a possibility of a hit in a search under each of the identification conditions with the character type unique to each of the identification conditions and the structure indicated by the search formula. claim 1 Symbol mounting structured document retrieval system, wherein the determining based on the information indicating the presence or absence of characters stored in the storage means. 3. The structured document search system according to claim 2, wherein the information indicating the presence or absence of characters is the number of characters.   Managing the character type information storage means, and when the structured document is stored in the database, for each structure of the structured document, identifying the character type for the text included in the structure and calculating the number of characters; And further comprising character type information management means for increasing the number of characters stored in the character type information storage means in association with the identified character type and the structure of the structured document by the calculated number of characters. Item 4. The structured document retrieval system according to item 3. A database for storing a plurality of structured documents and a same-view condition information storage unit for storing a plurality of same-view conditions for identifying the search keyword, and a structured document stored in the database as a target To the database server computer that performs the search process
Receiving a search request given from a client terminal;
Determining whether the search formula included in the accepted search request includes a batch identification condition for collectively specifying all the identification conditions;
If the search formula includes the batch identification condition, obtaining all of the plurality of identification conditions from the identification condition information storage unit;
Based on the presence / absence of characters corresponding to the structure indicated by the search expression among the characters included in the structured document stored in the database of the character types unique to each of the acquired identification conditions, Determining whether or not there is a possibility of a hit in each search;
Developing a search keyword included in the search formula based on the same viewing condition excluding the same viewing condition determined not to be a hit among the plurality of the same viewing conditions acquired; ,
Searching for a structured document including a structure that matches a search condition indicated by the search expression from the database based on the search keyword included in the search expression and the expanded search keyword. program.

Images