JP4388092B2 - Structured document database management system and program - Google Patents

Description

  The present invention relates to a structured document database management system that manages a structured document database in which a structured document such as an XML document is stored, and is particularly included in the structured document when the structured document is stored in the database. The present invention relates to a structured document database management system and program suitable for creating an index of words to be recorded.

  In a structured document, a hierarchical structure is expressed by a character string called a tag. As a representative of structured documents, for example, XML (Extensible Markup Language) format documents, that is, XML documents are widely known. XML has features such as hierarchization of data by means of tagged tags and free expandability of structure. An XML database (XMLDB) is known as an XML utilization technology that makes use of such features. The XML database is controlled by an XML database management system (XMLDBMS), thereby providing a function for storing and managing the hierarchical structure of XML data as it is, and a function for searching by a query language.

  In the XML database, a technique for constructing an index based on words included in the document is used in order to perform a high-speed search using words included in the XML document stored in the database (for example, Patent Document 1). reference). Such an index is generally constructed by analyzing the XML document and determining a word from the character sequence when the XML document is stored in the XML database.

For example, Patent Document 2 describes a technique for compressing a structured document by replacing a word character string extracted from the structured document with a shortened character string shorter than the word character string. In Patent Document 2, it is determined whether or not an input character is blank. If the character is blank, the character string stored in the memory until the blank is recognized is replaced with a replacement character, that is, the character string is determined for each blank. It is described that the structured document is shortened by dividing and replacing each delimited character string with a replacement character.
JP 2006-172268 A JP 2001-67348 A

  In the index construction technique (prior art) as described in Patent Document 1, a space (interval), a tab, a line feed character (LF (Line Feed)), and a carriage return character (CR (Carriage Return)) are blank. Characters are treated without distinction from other characters. In other words, in the prior art, the blank character is only one character constituting the text. For this reason, a word including a blank character is recognized only as a word including a blank character as it is.

  By the way, when a user creates an XML document, the XML document may be shaped into a form that is easy for humans to see on the display screen (or printing paper). For example, inserting a blank character between character strings of meaningful words is one form of document formatting. In the prior art, when such a word is selected as a search condition, the corresponding XML document is not searched.

  Therefore, it is conceivable to apply the technique described in the above cited document 2 and extract the word to be indexed by dividing the character string included in the structured document into blank characters. However, the blank character is not used only for formatting on the display, but may be necessary as an XML document. For this reason, the target XML document cannot always be searched by simply separating the blank characters.

  The present invention has been made in consideration of the above circumstances, and its purpose is to use words that make sense as a document by removing only white space characters for formatting among the white space characters included in structured documents when building an index. To provide a structured document database management system and program capable of constructing an index.

  According to one aspect of the present invention, a structured document database management system is provided. This structured document database management system analyzes the structure of a structured document to be stored in a structured document database, in which a hierarchical structure is expressed using tags, and detects the text that appears after the tag And a blank character string that does not make sense as a structured document that appears in the structured document analyzed by the syntax analyzing means is associated with structural information representing a structure associated with the blank character string. Blank character information accumulating means for accumulating; an ignorable blank character determining means for determining whether the text detected by the syntax analyzing means is an ignorable blank character composed of only blank characters; and the text is the ignorable If it is a whitespace character followed by a start tag followed by the relevant whitespace character, the relevant whitespace character is A first blank character judging means for storing in the blank character information accumulating means in association with the structure information representing the structure as the formatting blank character related to the structure of the next element including the tag, and the text Is not the ignorable blank character, the blank character string appearing in the text and the structure information of the structure related to the blank character string are compared with the information stored in the blank character information storage means, Second blank character determination means for determining whether the blank character string is a formatting blank character, and an index used for searching the structured document including the text are determined as the formatting blank character from the text. An index creating means for creating the data based on the text part from which the blank character string is removed and storing it in the structured document database, and the structured including the text Comprising a document storage processing means for storing the write to the structured document database.

  According to the present invention, from the text detected by analyzing the structure of the structured document, only the blank character determined to be the formatting blank character used for the formatting on the display is removed, and the text is included. By creating an index used to search for structured documents, words sandwiching white space for formatting can be registered in the index in the form of words that make sense as documents.

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 according to an embodiment of the present invention. The client-server system mainly includes a database server (database server computer) 10 and a plurality of client terminals. 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.

  The database server 10 has 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 an XML database management program 41 and an XML database 42.

  The XML database management program 41 is used for management of the XML database 42 by the database server 10 and search processing based on a search request from a client terminal. The XML database 42 stores a plurality of XML documents (XML format electronic documents) to be searched. In the present embodiment, the XML database 42 stores a set of XML documents (XML document set) 421 collected via the network 30. The XML database 42 also stores an index (index information) 422 that is used to search XML documents stored in the XML database 42.

  In the present embodiment, an XML database management system 50 is realized by the database server 10. FIG. 2 is a block diagram mainly showing a functional configuration of the XML database management system 50. The XML database management system 50 includes a document reading unit 51, a syntax analysis unit 52, a blank character determination unit 53, a blank character temporary storage unit 54, a blank character information storage unit 55, an index creation unit 56, and a document storage processing unit 57.

  The document reading unit 51 reads an XML document to be stored in the XML database 42. In the present embodiment, the document reading unit 51 reads an XML document that matches a predetermined condition from among XML documents published on the network 30. The syntax analysis unit 52 analyzes the XML structure (syntax) of the XML document read by the document reading unit 51.

  The blank character determination unit 53 determines the type of blank character that appears in the structure analyzed by the syntax analysis unit 52. In particular, the blank character determination unit 53 determines whether the blank character is a formatting blank character for formatting on display. The blank character determination unit 53 includes a first blank character determination unit 531, a second blank character determination unit 532, and a decomposition unit 533.

  The first white space character determination unit 531 is a white space character that appears in the structure analyzed by the syntax analysis unit 52 and is called “ignorable white space”, and has no meaning as an XML document. It is determined whether it is. The first white space character determination unit 531 determines that the white space character determined as “ignorable white space” is a white space character for formatting for display formatting, and determines the white space character (ignorable white space). The blank character is stored in the blank character temporary storage unit 54 in association with the structure information of the structure including the element content.

  The first blank character determination unit 531 has a start tag next to the blank character (blank character string) that is determined to be a formatting blank character (formatting blank character string) and stored in the temporary blank character storage unit 54. When it appears, the blank character is stored in the blank character information storage unit 55 as the formatting blank character related to the next element (structure) including the start tag in association with the structure information of the next element.

  When the text is analyzed by the syntax analysis unit 52, the decomposition unit 533 decomposes the text into text that does not include a blank character (blank character string) and a blank character (blank character string).

  The second blank character determination unit 532 has a position of a blank character (blank character string) in the text decomposed by the decomposition unit 533, a relationship with a preceding blank character (blank character string), and a structure including the text. Based on the relationship with the string length of the formatting blank character (formatting blank character string) related to the structure one level above, the blank character (blank character string) is the formatting blank character (formatting blank character string). It is determined whether it is a candidate for white space for formatting (white space character string for formatting) or just a white space character (blank character string) that appears in the text.

  The second blank character determination unit 532 has a structure in which a blank character (blank character string) in the text appears at the head of the text, and the character string length of the blank character (blank character string) includes the text. Formatting space related to the structure (corresponding structure) that includes the text if the character string length of the formatting blank character (formatting blank character string) related to the upper structure is longer than the character string length It is determined as a character (blank string for formatting). In this case, the second blank character determination unit 532 associates the determined blank character (blank character string) with the structure information of the corresponding structure as the formatting blank character (formatting blank character string), and stores the blank character information storage unit. Accumulate in 55.

  The position of the first blank character (blank character string) that appears in the text is not at the beginning or end of the text, and the character string length of the blank character (blank character string) is one level higher in the structure including the text. If the character string length of the formatting blank character (formatting blank character string) associated with the structure is longer than the character string length, the second blank character determination unit 532 uses the blank character (blank character string) for formatting associated with the corresponding structure. It is determined as a candidate for a blank character (formatting blank character string). In this case, the second blank character determination unit 532 associates the determined blank character (blank character string) with the structure information of the corresponding structure as a candidate for the blank character for formatting (blank character string for formatting), and blank character information. Accumulate in the accumulation unit 55.

  The second blank character determination unit 532 determines a blank character (blank character string) determined as a formatting blank character (formatting blank character string) or determined as a candidate for a formatting blank character (formatting blank character string). ) If the same white space character (blank character string) as the white space character (blank character string) appears later than the white space character (blank character string) in the text containing (Character string) is determined as a formatting blank character (formatting blank character string), or a candidate for a formatting blank character (formatting blank character string) is determined as a formatting blank character (formatting blank character string).

  The blank character temporary storage unit 54 is used to temporarily store the blank character string determined by the blank character determination unit 53 as a formatting blank character in association with the structure information of the corresponding structure. For example, a first specific area on the memory 11 of the database server 10 of FIG. 1 is allocated to the blank character temporary storage unit 54.

  The blank character information storage unit 55 determines a blank character string determined by the blank character determination unit 53 as a candidate for a formatting blank character or a formatting blank character (that is, a blank character string that does not make sense as an XML document). It is used for storing in association with the structure information representing the structure related to the blank character string. For example, a second specific area on the memory 11 is allocated to the blank character information storage unit 55.

  The index creation unit 56 searches the original XML document (the original XML document) based on the XML document in which the syntax is analyzed by the syntax analysis unit 52 and the blank character for formatting (the blank character string determined as) is removed. Create an index (index record) that is used to The document storage processing unit 57 stores (registers) the XML document original text in the XML database 42. In the XML document stored in the XML database 42, the character strings before and after the blank character are treated as being continuous.

  In the present embodiment, each function unit (processing unit) of the document reading unit 51, the syntax analysis unit 52, the blank character determination unit 53, the index creation unit 56, and the document storage processing unit 57 is stored in the database server 10 of FIG. The XML database management program 41 stored in the device 40 is assumed to be realized by reading it into a memory (not shown) 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.

  FIG. 3 shows an example of the data structure of the index 422 stored in the XML database 42 in association with the XML documents in the XML document set 421. The index 422 is stored in the XML database 42 using, for example, BTree. In the index 422, character strings (index character strings) are arranged in ascending order according to the order of the character strings, and are divided and stored in a plurality of areas called pages. A page is a unit for reading / writing data from / to the external storage device 40.

  In FIG. 3, one of index page groups constituting the index 422 is shown as an index page 422a. The index page 422a is composed of a set of records (index records) having information on the appearance position of the keyword for each keyword (character string constituting the keyword). For example, the ID of the XML document (XML document original text) in which the corresponding keyword appears is used as the appearance position information.

  Next, the parsing process including the blank character determination for formatting, which is executed when the XML document is stored in the XML database 42 in the XML database management system 50 of FIG. 2, will be described with reference to the flowcharts of FIGS. 4A and 4B. To do.

  The document reading unit 51 stores, in the XML database 42, an XML document (target XML document) that matches a predetermined condition among XML documents published on the network 30 by a Web server or a file server. As a power XML document, for example, document reading (step S1) to be read into the memory 11 (memory 11 in the XML database management system 50) of the database server 10 of FIG. 1 is started.

FIG. 5 shows an example of an XML document (first XML document) that has been read by the document reading unit 51.
As is well known, an XML document (structured document) is configured in units of elements. An element is a name that represents the contents of data and is sandwiched between tags. This name (namely, element name) is described in the tag. In an XML document, a so-called nested structure in which an element is included in an element can be used.

  In the example of the XML document shown in FIG. 5, the entire document is an element sandwiched between <patent> tags, and an “information processing apparatus” sandwiched between <name> tags indicating the name of the document includes the contents of the document. It shows that it is the name of one patent.

  In the XML document shown in FIG. 5, for example, the start position of the <name> tag and the <summary> tag (position of the start tag) is set so that a person can easily see the document while the XML document is displayed on the screen. It is described with a tab character one level lower than the starting position of the <patent> tag, which is the top of the structure. In addition, the text in the <summary> tag, that is, the text (text) that is the content (value) of the element (summary element) containing the <summary> tag, is inserted into the <summary> tag. It is described with one step lower than that. As described above, the XML document shown in FIG. 5 includes a blank character.

In order to make it easy to understand the existence of these white space characters (that is, a line feed character and a tab character), for example, if “↓” indicates a line feed character and “t” indicates one tab character, the XML shown in FIG. When the document is processed by the XML database management system 50 (database server 10), the following document “<patent> ↓ t <name> information processing apparatus </ name> ↓ t <summary> ↓ tt The document is read, the structure is analyzed, and the extracted structure information ↓ tt and the original text are stored in the external storage device.The user can extract the necessary information ↓ tt by keyword ↓ t </ summary > ↓ </ patent>
Are treated as
FIG. 6 shows the XML document shown in FIG. 5 in a format in which the line feed character and the tab character included in the XML document are replaced with symbols “↓” and “t”.

  Now, the document reading unit 51 reaches the end of the XML document, and there is no data to be read in the document reading (step S1) for the XML document (target XML document) shown in FIG. 6 (FIG. 5). The process is repeated until the XML document ends.

  On the other hand, the syntax analysis unit 52 analyzes the structure of the document data read into the memory 11 by the document reading unit 51 until the target document ends (step S2) (step S3).

  Generally, in order to take out words in an XML document and create an index (index record), it is necessary to combine adjacent characters and recognize them as words. However, in the example of the XML document described above, a blank character (new line character “↓”) and a character “information” included in “structure information” (between the characters “information” and “information”) Two tab characters “t”) are inserted. For this reason, for example, the word “structure information” (or the word “information”) is not extracted as an index. Therefore, in this embodiment, as described below, it is possible to extract “structure information” as an index from the character string “structure information ↓ tt information”.

  The syntax analysis unit 52 is, for example, an XML parser having an XML document analysis function. In the present embodiment, the syntax analysis unit 52 imports and analyzes the XML document read into the memory 11 by the document reading unit 51 using a known application interface called SAX (Simple API for XML) 2. In this case, the syntax analysis unit 52 is notified in units of element start (start tag), element end (end tag), character string (text), and “ignorable white space”. In accordance with this notification, the syntax analysis unit 52 performs syntax analysis (structural analysis) on the XML document (target XML document) shown in FIG. 5 in units of start tag, end tag, text, and “ignorable white space”. ) Is repeated until the end of the XML document (steps S2 and S3).

In the example of the XML document in FIG. 6, the XML document is document data shown in the following order.
(1) <Patent>: Start tag indicating the element of the highest level (first level) structure (hierarchical structure)
(2) ↓ t: The content of the <patent> element, but “ignorable white space”
(3) <Name>: Start tag indicating the element of the second level structure
(4) Information processing device: Text that is the contents of the <name> element
(5) </ Name>: End tag that indicates the end of the element of the second level structure
(6) ↓ t: The content of the <patent> element, but “ignorable white space”
(7) <Summary>: Start tag indicating the element of the second level structure
(8) ↓ tt This device reads a document, analyzes its structure, and stores both the extracted structure information ↓ tt information and the original text in an external storage device. The user can take out the necessary information ↓ tt by keyword. ↓ t
: <Summary> Text that is the content of the element (hereinafter referred to as text T)
(9) </ Summary>: End tag that indicates the end of the element in the second stage structure
(10) ↓: The content of the <patent> element is "ignorable white space"
(11) </ patent>: Decomposed into an end tag indicating the end of the element of the topmost structure. Here, a description of the unit corresponding to the data is attached to the right side of the document data.

  The document storage processing unit 57 stores the document data decomposed by the syntax analysis unit 52 in the XML database 42 as a part of the XML document set 421 as the original text read by the document reading unit 51.

  In an XML document, as is well known, as a white space character (white space symbol) called “white space”, a space (half-width space), a tab, a line feed character (LF (Line Feed)), and a return character (CR (Carriage Return)) Is defined. A line feed character and a carriage return character are blank characters representing a line feed.

  The above-described white space is generally used by being inserted between distinguishing elements such as separating element names and attribute names in tags. However, in order to make it easy to distinguish between tags or to make them easier to read, a blank character may be inserted for formatting. In XML, regardless of the XML structure, white space characters that are not meaningful text data (indicating the content that is the component of the element) are defined as “ignorable white space”. Yes. The “ignorable white space” is text data usually composed of only white space characters sandwiched between a tag (end tag) and a tag (start tag).

  In the present embodiment, the blank character determination unit 53 acquires the result of syntax analysis by the syntax analysis unit 52 and determines a blank character (particularly, a formatting blank character) based on the structure analysis result. The blank character determination unit 53 stores the blank character determined to be the formatting blank character in the blank character information storage unit 55 in association with the structure information indicating the structure related to the blank character.

  In the present embodiment, the following two types are treated as formatting blank characters (or formatting blank character candidates).

  The first is “ignorable white space”. More specifically, when the structure following “ignorable white space” is a start tag, the formatting white space character for the structure represented by the start tag (the structure of the element including the start tag) is immediately before it. The “ignorable white space” that appeared in

  Second, if a blank character string exists in the text and is a blank character string longer than a formatting blank character string for a structure one level higher than the structure indicating the text, the formatting for the structure indicating the text is performed. The blank character for use is a blank character string existing in the text.

Hereinafter, the details of the processing including the determination of the blank character for formatting based on the result of the syntax analysis by the syntax analysis unit 52 will be described in the XML document (first XML document) in FIG. 6 in the above document data (1) to (11). A case where the analysis is performed after being decomposed will be sequentially described with reference to FIGS. 7 and 8 according to the flowcharts of FIGS. 4A and 4B. FIG. 7 is a diagram showing changes in the contents of the blank character information storage unit 55, and FIG. 8 is a diagram showing changes in the contents of the blank character temporary storage unit 54. In FIG. 7 and FIG. 8 and the following description, the structure information representing the analyzed structure is represented in the XPath format. As structure information, an ID (node ID) unique to an element (node) of the corresponding structure can be used.
(1) <patent> (/ patent)
The <patent> tag at the top of the XML document is a start tag indicating the start of the <patent> element indicated by the structure information “/ patent” (steps S4 and S5). In this case, if there is a formatting blank character preceding the structure information “/ patent”, the first blank character determining unit 531 in the blank character determining unit 53 converts the structure information “/ patent” into the formatting blank character. Are stored in the blank character information storage unit 55 in association with (step S6). However, in the present embodiment, the structure information “/ patent” is the head of the line, and there is no formatting blank character preceding the structure information “/ patent”. In such a case, the first blank character determination unit 531 accumulates only the structure information “/ patent” in the leading entry of the blank character information accumulation unit 55.

  FIG. 7A shows the contents of the blank character information storage unit 55 at this time. The entry of the blank character information storage unit 55 stores a flag (flag information) in pairs with the structure information and the formatting blank character (the blank character determined to be). The flag in this entry indicates whether or not the blank character (blank character string) accumulated in the entry is confirmed as a formatting blank character (formatting blank character string) (whether it is a candidate for a formatting blank character). ). However, in the present embodiment, as in the example of FIG. 7A, a flag in an entry that does not have a formatting blank character is set to a state indicating “determined” for convenience.

(2) ↓ t
When the structure information “/ patent” is accumulated in the blank character information accumulation unit 55 (step S6), the syntax analysis unit 52 analyzes the structure of the document data following the <patent> tag (steps S2 and S3). The document data following the <patent> tag is “↓ t”. This “↓ t” is (part of) the content of the <patent> element and is “ignorable white space” (steps S4 and S8).

  In this case, the first blank character determination unit 531 determines that “↓ t” is a formatting blank character (formatting blank character string) related to the element that appears next. Therefore, the first blank character determination unit 531 uses <↓> as a formatting blank character, and <patent> element structure information (currently a processing target) with the content of (↓ t) as “↓ t”. The information is stored in the blank character temporary storage unit 54 in association with “/ patent” (step S9). FIG. 8A shows the contents of the blank character temporary storage unit 54 at this time. Even if the determination in step S8 (determination of “ignorable white space”) is performed by the first blank character determination unit 531, the determination is independent from the syntax analysis unit 52 and the first blank character determination unit 531. Part ("ignorable white space" determination part).

(3) <Name> (/ Patent / Name)
When the formatting blank character “↓ t” is associated with the structure information “/ patent” and stored in the temporary blank character storage unit 54 (step S9), the syntax analysis unit 52 stores the document data subsequent to “↓ t”. The structural analysis is performed (steps S2 and S3). The document data following “↓ t” is a <name> tag.

  The <name> tag is a start tag indicating the start of the <name> element indicated by the structure information “/ patent / name” (steps S4 and S5). In this case, the first blank character determination unit 531 refers to the blank character temporary storage unit 54, so that the formatting blank character (in this case, “/ patent / name” is the document data obtained previously). ↓ t ”), that is, that there is a formatting blank character (“ ↓ t ”) preceding the structure information“ / patent / name ”. Therefore, the first blank character determination unit 531 displays the structure information “/ patent / name” to indicate that the structure information “/ patent / name” is described subsequent to the previously obtained blank character for formatting “↓ t”. “/ Patent / name” is associated with the preceding formatting blank character “↓ t” and stored in the second entry of the blank character information storage unit 55 (step S6). At this time, the first blank character determination unit 531 sets the flag of the second entry of the blank character information storage unit 55 in which the formatting blank character “↓ t” is stored to a state indicating “confirmed”. FIG. 7B shows the contents of the blank character information storage unit 55 at this time.

(4) Information processing device (text)
The document data following the <name> tag is the text “information processing apparatus” (steps S4 and S8). In this case, the decomposition unit 533 in the blank character determination unit 53 decomposes the text into a blank character string and a text (text part) not including the blank character string in order from the top (steps S10 and S11).

  Here, the text “information processing apparatus” does not include a blank character string (steps S12 and S13). In this case, the second blank character determining unit 532 in the blank character determining unit 53 determines that the entire “information processing apparatus” is meaningful text (step S14). It is clear that the text “information processing apparatus” determined in step S14 does not include a formatting blank character. The second blank character determination unit 532 is a text storage unit (not shown) in which the meaningful information determined in step S14 (that is, text that does not include a blank character string) “information processing apparatus” is secured in the memory 11. To store.

(5) </ Name> (End of "/ Patent / Name")
</ Name> is an end tag indicating the end of the <name> element indicated by the structure information “/ patent / name” of the second level hierarchy (steps S4 and S5). In this case, the syntax analysis unit 52 raises the processing target hierarchy (target hierarchy) by one level (step S7). As a result, the structure information of the hierarchy to be processed becomes “/ patent”. Then, the syntax analysis unit 52 analyzes the structure of the document data following the </ name> tag (steps S2 and S3).

(6) ↓ t
The document data following the </ name> tag is “↓ t”. This “↓ t” is the content of (part of) the <patent> element constituting the current processing target hierarchy, and is “ignorable white space” (steps S4 and S8). Therefore, the first blank character determination unit 531 assumes that “ignorable white space”, that is, “↓ t” is a blank character for formatting the element that appears next, and regards “↓ t” as the structure of the current processing target hierarchy. It is stored in the blank character temporary storage unit 54 in association with the information “/ patent” (step S9). The contents of the blank character temporary storage unit 54 at this time are the same as those in FIG.

(7) <Abstract> (/ Patent / Abstract)
The document data following the “↓ t” (ignorable white space) determined to be a formatting blank character is a <summary> tag. The <summary> tag is a start tag indicating the start of the <summary> element indicated by the structure information "/ patent / abstract" (steps S4 and S5).

  In this case, the first blank character determination unit 531 refers to the blank character temporary storage unit 54, so that the formatting blank character (in this case, “/ patent / abstract” is the document data obtained last time. ↓ t ”) is recognized following the description. Therefore, the first blank character determination unit 531 displays the structure information “/ patent / abstract” to indicate that the structure information “/ patent / abstract” is described subsequent to the previously obtained blank character for formatting “↓ t”. “/ Patent / abstract” is associated with the preceding formatting blank character “↓ t” and stored in the third entry of the blank character information storage unit 55 (step S6). At this time, the first blank character determination unit 531 sets the flag of the third entry of the blank character information storage unit 55 in which the formatting blank character “↓ t” is stored to a state indicating “confirmed”. FIG. 7C shows the contents of the blank character information storage unit 55 at this time.

(8) T (text)
<Summary> The document data following the tag is text T, that is, "↓ tt This device reads the document, analyzes its structure, and stores the retrieved structure information ↓ tt information and original text in an external storage device. The user can take out the necessary information ↓ tt by keyword. ↓ t "
(Steps S4 and S8). In this case, the decomposition unit 533 decomposes the text T in order from the top into a blank character string and a text portion that does not include the blank character string (hereinafter referred to as meaningful text) (steps S10 and S11).

Here, the text T is
8-1) Blank character string “↓ tt”
8-2) Meaningful text “This device reads the document, analyzes its structure, and extracts it” (hereinafter referred to as meaningful text T1)
8-3) Blank character string “↓ tt”
8-4) Meaningful text “Both information and original text are stored in an external storage device. The user uses a keyword” (hereinafter referred to as meaningful text T2).
8-5) Blank character string “↓ tt”
8-6) Meaningful text “Necessary information can be extracted” (hereinafter referred to as meaningful text T3)
8-7) Blank character string “↓ t”
Is broken down into

  Hereinafter, processing for each data of the above-described 8-1) to 8-7) in which the text T is decomposed will be sequentially described.

8-1) Blank character string “↓ tt”
This blank character string “↓ tt” appears at the beginning of the text T (step S15). In addition, this blank character string “↓ tt” includes the structure information “/ patent / abstract” representing the structure one level higher than the structure information “/ patent / abstract / text ()” representing the structure including the text T and the blank character. The character string length is longer than the shaping blank character “↓ t” (see FIG. 7C) associated with the third entry in the information storage unit 55. In this case, the second blank character determination unit 532 determines that the blank character string “↓ tt” is a blank character for formatting the text T (step S16). In step S16, the second blank character determination unit 532 associates “↓ tt” determined as the formatting blank character with the structure information “/ patent / abstract / text ()” representing the structure including the text T. Thus, it is stored in the blank character information storage unit 55. FIG. 7D shows the contents of the blank character information storage unit 55 at this time.

8-2) Meaningful text T1
T1 following the first blank character string “↓ tt” of the text T is not a blank character string (step S13). In this case, the second blank character determination unit 532 determines that “T1 is a meaningful text (step S14). The second blank character determination unit 532 determines the meaningful text determined in step S14. For example, T1 is stored behind the already stored text (“information processing apparatus”) in the text storage unit.

8-3) Blank character string “↓ tt”
The blank character string “↓ tt” (that is, the second blank character string in the text T) following the meaningful text T1 appears at a position other than the beginning and end of the text T (that is, the middle part of the text T) (step S13). , S15, S17). As described above, when the blank character string “↓ tt” appears in the middle portion of the text T, the second blank character determination unit 532 refers to the blank character information storage unit 55 and includes the structure including the text T (corresponding structure). ) Whether or not formatting blank characters (formatting blank character strings) are stored in association with the structure information “/ patent / abstract / text ()” (that is, formatting blank characters of the corresponding structure are stored) Is determined (step S19). At this time, in the fourth entry of the blank character information storage unit 55, the blank character string “↓ tt” at the beginning of the text T is used as a fixed blank character string for formatting as shown in FIG. It is stored in association with the structure information “/ patent / abstract / text ()”.

  The second blank character determination unit 532 determines that the blank character string “↓ tt” currently being processed when the blank character string for formatting with the corresponding structure is accumulated in the blank character information accumulation unit 55 (step S19). Is matched with the formatting blank character (here, the formatting blank character string “↓ tt”) (step S21).

As in the present embodiment, when the blank character string “↓ tt” currently being processed matches the formatting blank character string of the corresponding structure stored in the blank character information storage unit 55, the second blank The character determination unit 532 determines whether or not the shaping blank character string is a candidate (has been confirmed) (step S23). Here, the formatting blank character string “↓ tt” of the corresponding structure, that is, the leading blank character string “↓ tt” of the text T has been confirmed as shown in FIG. In this case, the second blank character determination unit 532 accumulates the blank character string “↓ tt” (that is, the second blank character string in the text T) currently being processed in the blank character information accumulation unit 55. As with the first (first) blank character string “↓ tt” in the text T, it is determined to be a formatting blank character related to the structure information “/ patent / abstract / English / text ()”. (Step S25)
8-4) Meaningful text T2 following the second blank character string “↓ tt” in the text T is not a blank character string (step S13). In this case, T2 is determined to be a meaningful text in the same manner as T1 (step S14). The second blank character determination unit 532 stores the meaningful text T2 determined in step S14 in the text storage unit, for example, following the text T1 already stored in the text storage unit. .

8-5) Blank character string “↓ tt”
The blank character string “↓ tt” following the meaningful text T2, that is, the third blank character string “↓ tt” in the text T appears in the middle part of the text T (steps S13, S15, S17). In this case, the third blank character string “↓ tt” is processed in the same manner as the second blank character string “↓ tt”.

8-6) Meaningful text T3
T3 subsequent to the third blank character string “↓ tt” in the text T is not a blank character string (step S13). In this case, T3 is determined to be a meaningful text in the same manner as T1 (step S14). The text T3 is stored in the text storage unit so as to follow the text T2 already stored in the text storage unit.

8-7) Blank character string “↓ t”
The blank character string “↓ t” following the meaningful text T3, that is, the fourth blank character string “↓ t” in the text T appears at the end of the text T (steps S13, S15, and S17). In this case, the second blank character determination unit 532 determines that the last (fourth) blank character string “↓ t” of the text T is a formatting blank character string related to the next element, and The character string “↓ t” is stored in the blank character temporary storage unit 54 in association with the structure information “/ patent / abstract / text ()” of the corresponding structure (step S18). The document data next to the blank character string “↓ t” (blank character string at the end of the text T) stored in the blank character temporary storage unit 54 is an end tag of the <summary> element, and the next element does not exist. The blank character string “↓ t” is not used as a formatting blank character string.
Thus, the process for the text T is completed.

(9) </ Summary>
When the process for the text T is completed, the syntax analysis unit 52 performs a structure analysis of document data subsequent to the text T (steps S2 and S3). The document data following the text T is a </ summary> tag.

  The </ summary> tag is an end tag indicating the end of the <summary> element indicated by the structure information "/ patent / summary" (steps S4 and S5). In this case, the syntax analysis unit 52 raises the target hierarchy by one level (step S7). As a result, the structure information of the hierarchy to be processed becomes “/ patent”. Then, the syntax analysis unit 52 analyzes the structure of the document data following the </ summary> tag (steps S2 and S3).

(10) ↓
</ Summary> The document data following the tag is “↓”. This “↓” is the content of (part of) the <patent> element constituting the current processing target hierarchy, and is “ignorable white space” (steps S4 and S8). Therefore, the first blank character determination unit 531 assumes that “ignorable white space”, that is, “↓” is the white space for formatting the element that appears next, and regards the “↓” as the structure information “ / Patent ”is stored in the blank character temporary storage unit 54 (step S9). FIG. 8B shows the contents of the blank character temporary storage unit 54 at this time. Since the document data next to “↓” (ignorable white space) stored in the temporary space character storage unit 54 is an end tag of the <patent> element, and the next element does not exist, the “↓” is a formatting blank. Not used as a string.

(11) </ patent>
When “↓” that is “ignorable white space” following the </ element> tag is processed, the syntax analysis unit 52 analyzes the structure of the document data following the “↓” (steps S2 and S3). Document data following this “↓” (ignorable white space) is a </ patent> tag. The </ patent> tag is an end tag indicating the end of the <patent> element in the highest hierarchy indicated by the structure information “/ patent” (steps S4 and S5). In this case, the syntax analysis unit 52 determines the end of the structure information “/ patent” (step S2), and ends the analysis of the XML document.

  At this time, the text “information processing apparatus” is stored in the text storage unit. The text storage unit further stores T1, T2, and T3 for the text T. That is, in the text storage unit, in addition to the text “information processing apparatus”, a part obtained by removing the formatting blank character string from the text T is stored.

  The index creation unit 56 is based on the text “information processing apparatus” stored in the text storage unit and the parts (T1, T2, and T3) from which the formatting blank character string is removed from the text T, that is, Based on a certain text, an index (index record) used to search for an XML document including the meaningful text is created.

  The index creation unit 56 stores the created index record in the XML database 42 as a part of the index 422 in association with the XML document original. In this case, since “structure information ↓ tt information” included in the text T is recognized as the word “structure information”, an index record having “structure information” as a keyword can be created.

  Next, it is assumed that an XML document (second XML document) different from the XML document (first XML document) shown in FIG. 5 is read by the document reading unit 51. FIG. 9 shows an example of the second XML document. The difference between the second XML document shown in FIG. 9 and the first XML document shown in FIG. 9 is that the text contains a blank character, but the text does not contain a blank character at the beginning. And an element appears in the text.

If the second XML document shown in FIG. 9 indicates a line feed character with “↓”, a tab character with “t”, and a space character with “s”, this document is processed by the XML database management system 50. The following document “<patent> ↓ t <name> information processing device </ name> ↓ t <summary> This device reads a document, analyzes its structure, extracts the structure information ↓ tsssss Both the information and the original text are stored in the external storage device.The user can retrieve the necessary information ↓ tsssss by keyword ↓ tt <English> ↓ tt The device reads a document and stores both the structure information ↓ ttand texts in an external storage. ↓ tt </ English> ↓ t </ abstract> ↓ </ patent>"
Are treated as
FIG. 10 shows the XML document shown in FIG. 9 in a format in which the line feed character and the tab character included in the XML document are replaced with symbols “↓” and “t”.

In the example of the XML document (second XML document) in FIG. 10, the XML document is document data organized and shown below.
(1) <Patent>: Start tag indicating the element of the highest structure
(2) ↓ t: The content of the <patent> element, but “ignorable white space”
(3) <Name>: Start tag indicating the element of the second level structure
(4) Information processing device: Text that is the contents of the <name> element
(5) </ Name>: End tag that indicates the end of the element of the second level structure
(6) ↓ t: The content of the <patent> element, but “ignorable white space”
(7) <Summary>: Start tag indicating the element of the second level structure
(8) This apparatus reads a document, analyzes its structure, and stores the extracted structure information ↓ tssssss and original text in an external storage device. The user can take out necessary information by using a keyword. ↓ tt
: <Summary> Text that is the content of the element (hereinafter referred to as text Ta)
(9) <English>: Start tag indicating the element of the third level structure
(10) ↓ ttThe device reads a document and stores both the structure information ↓ ttand texts in an external storage.
: Text that is the content of the <English> element (hereinafter referred to as text Tb)
(11) </ English>: End tag that indicates the end of the element of the third level structure
(12) ↓ t: <element> The contents of the element, but “ignorable white space”
(13) </ Summary>: End tag that indicates the end of the element in the second stage structure
(14) ↓: <patent> element content, but "ignorable white space"
(15) </ patent>: Decomposed into an end tag indicating the end of the element of the highest hierarchy.

  Hereinafter, in the case where the XML document (second XML document) in FIG. 10 is analyzed after being decomposed into the above document data (1) to (15), according to the flowcharts in FIGS. 4A and 4B, FIGS. Will be described sequentially. FIG. 11 is a diagram showing changes in the contents of the blank character information storage unit 55, and FIG. 12 is a diagram showing changes in the contents of the blank character temporary storage unit 54.

  Of the document data (1) to (15) in the XML document (second XML document) of FIG. 10, the content of the processing for the document data (1) to (7) is the XML document (first XML document of FIG. 6). This is the same as the content of the processing for the document data (1) to (7) in (Document). Therefore, here, processing for document data (8) to (15) in the XML document (second XML document) of FIG. 10 will be described. The processing for the document data (1) to (7) in the XML document (second XML document) in FIG. 10 is performed on the document data (1) to (7) in the XML document (first XML document) in FIG. Refer to the explanation on processing.

(8) Ta (text)
<Summary> The document data Ta following the tag is the text of the structure information “/ patent / abstract”, that is, “This device reads the document, analyzes the structure, extracts the structure information ↓ tssssss and the original text. Both are stored in the external storage device.The user can retrieve necessary information by using the keyword ↓ tsssss. ↓ tt ”
(Steps S4 and S8). In this case, the decomposition unit 533 decomposes the text Ta into a blank character string and a text part (meaningful text) that does not include the blank character string in order from the top (steps S10 and S11).

Here, the text Ta is
8-1) Meaningful text “This device reads the document, analyzes its structure, and extracts it” (hereinafter referred to as meaningful text Ta1)
8-2) Blank character string “↓ tsssss”
8-3) Meaningful text “Both information and original text are stored in an external storage device. The user uses a keyword” (hereinafter referred to as meaningful text Ta2).
8-4) Blank character string “↓ tsssss”
8-5) Meaningful text “Necessary information can be retrieved” (hereinafter referred to as meaningful text Ta3)
8-6) Blank string “↓ tt”
Is broken down into

  Hereinafter, processing for each data of the above-described 8-1) to 8-6) in which the text Ta (that is, the text of the structure information “/ patent / abstract”) is decomposed will be sequentially described.

8-1) Meaningful text Ta1
Since the beginning of the text T is a meaningful text Ta1 and not a blank character string (steps S13 and S14), the structure of the next element “/ patent / abstract / text ()” is obtained when the text Ta1 is analyzed. The formatting whitespace associated with is not yet determined. The text Ta1 is stored in the text storage unit.

8-2) Blank character string “↓ tsssss”
The blank character string “↓ tsssss” following the meaningful text Ta1 appears in the middle of the text Ta1 (steps S13, S15, S17). This blank character string “↓ tsssss” is a blank character string that appears first (first) in the text Ta1, and is a formatting blank related to the structure information “/ patent / abstract / text ()” corresponding to the text Ta1. The character does not exist in the blank character information storage unit 55 (step S19). The blank character string “↓ tssssss” is structural information “/ patent / abstract” representing the structure one level higher than “/ patent / abstract / text ()” and the third entry in the blank character information storage unit 55. The character string length is longer than the corresponding shaping blank character “↓ t” (see FIG. 7C).

  In this case, the second blank character determination unit 532 uses the blank character string “↓ tssssss” as a candidate for blank characters for formatting related to the structure information “/ patent / abstract / text ()”. In association with “patent / abstract / text ()”, it is stored in the fourth entry of the blank character information storage unit 55 in the state of FIG. 7C (step S 20).

  FIG. 11A shows the contents of the blank character information storage unit 55 at this time. In FIG. 11A, the flag included in the fourth entry of the blank character information storage unit 55 is set to a state indicating “candidate (candidate blank character for shaping)”.

8-3) Meaningful text Ta2
Ta2 following the first blank character string “↓ tsssss” in the text Ta is not a blank character string (step S13). In this case, Ta2 is determined to be a meaningful text (step S14). The text Ta2 is stored in the text storage unit so as to follow the text Ta1 already stored in the text storage unit.

8-4) Blank character string “↓ tsssss”
The blank character string “↓ tsssss” following the meaningful text Ta1, that is, the second blank character string “↓ tsssss” in the text Ta appears in the middle part of the text Ta (steps S13, S15, S17). This second blank character string “↓ tssssss” matches the first blank character string “↓ tssssss” stored as a candidate for the formatting blank character in the fourth entry of the blank character information storage unit 55 ( Steps S19, S21, S23). In this case, the second blank character determination unit 532 determines that both the first blank character string and the second blank character string (that is, the blank character string currently being processed) are formatting blank characters. The first blank character string “↓ tssssss” stored in the fourth entry of the character information storage unit 55 is determined as a formatting blank character string (step S24). That is, the second blank character determination unit 532 sets the flag included in the fourth entry of the blank character information storage unit 55 in which the blank character string “↓ tsssss” is stored as “candidate” in FIG. The state shown is changed to the state showing “confirmed”. FIG. 11B shows the contents of the blank character information storage unit 55 at this time.

8-5) Meaningful text Ta3
Ta3 following the second blank character string “↓ tsssss” in the text Ta is not a blank character string (step S13). In this case, Ta3 is determined to be a meaningful text (step S14). The text Ta3 is stored in the text storage unit so as to follow the text Ta2 already stored in the text storage unit.

8-6) Blank string “↓ tt”
The blank character string “↓ tt” following the meaningful text Ta3, that is, the third blank character string “↓ tt” in the text Ta appears at the end of the text Ta (steps S13, S15, S17). In this case, the second blank character determination unit 532 determines that the last (fourth) blank character string “↓ tt” of the text Ta is a formatting blank character string related to the next element, and The character string “↓ tt” is stored in the blank character temporary storage unit 54 in association with the structure information “/ patent / abstract /” of the corresponding structure (step S18). FIG. 12A shows the contents of the blank character temporary storage unit 54 at this time.

  Thus, the process for the text Ta is completed.

(9) <English> (/ Patent / Abstract / English)
The document data following the blank character string “↓ tt” (determined as a blank character string for formatting) at the end of the text Ta is an <English sentence> tag. The <English> tag is a start tag indicating the start of the <English> element indicated by the structure information “/ patent / abstract / English” (steps S4 and S5). In this case, the first blank character determination unit 531 refers to the blank character temporary storage unit 54, so that the formatting blank character string (structural information “/ patent / abstract / English”) is the previously acquired document data ( Here, it is recognized that it is described subsequent to “↓ tt”). Therefore, the first blank character determination unit 531 indicates that the structure information “/ patent / abstract / English” is described subsequent to the previously obtained blank character string “↓ tt” for formatting. The structure information “/ patent / name / English” is stored in the fifth entry of the blank character information storage unit 55 in association with the preceding blank character string “↓ tt” (step S6). At this time, the first blank character determination unit 531 sets the flag of the fifth entry of the blank character information storage unit 55 in which the formatting blank character “↓ tt” is stored to a state indicating “confirmed”. FIG. 11C shows the contents of the blank character information storage unit 55 at this time.

(10) Tb (text)
The document data following the <English> tag is the text Tb, that is, “↓ tt The device reads a document and stores both the structure information ↓ ttand texts in an external storage.
(Steps S4 and S8). In this case, the decomposition unit 533 decomposes the text Tb in order from the top into a blank character string and a text portion (a meaningful text) that does not include the blank character string (steps S10 and S11).

Here, the text Tb is
10-1) Blank character string “↓ ttt”
10-2) Meaningful text “The device reads a document and stores both the structure information” (hereinafter referred to as meaningful text Tb1)
10-3) Blank character string “↓ tt”
10-4) Meaningful text “and texts in an external storage” (hereinafter referred to as meaningful text Tb2)
10-5) Blank character string “↓ tt”
Is broken down into

  Hereinafter, processing for each of the above data 10-1) to 10-5) in which the text Tb (that is, the text of the structure information “/ patent / abstract / English”) is decomposed will be sequentially described.

10-1) Blank character string “↓ ttt”
This blank character string “↓ ttt” appears at the beginning of the text Tb (step S15). This blank character string “↓ tt” is the structure information “/ patent / abstract / English” representing the structure one level higher than the structure information “/ patent / abstract / English / text ()” representing the structure including the text Tb. "Is longer than the formatting blank character" ↓ tt "(see FIG. 11C) associated with the fifth entry in the blank character information storage unit 55.

  In this case, the second blank character determination unit 532 determines that the blank character string “↓ ttt” is a blank character for formatting the text Tb (step S16). In step S16, the second blank character determination unit 532 converts “↓ ttt” determined as the formatting blank character into the structure information “/ patent / abstract / English / text ()” representing the structure including the text Tb. Correspondingly, it is stored in the sixth entry of the blank character information storage unit 55. FIG. 11D shows the contents of the blank character information storage unit 55 at this time.

10-2) Meaningful text Tb1
Tb1 following the first blank character string “↓ ttt” of the text Tb is not a blank character string (step S13). In this case, the second blank character determination unit 532 determines that Tb1 is a meaningful text (step S14). The text Tb1 is stored in the text storage unit.

10-3) Blank character string “↓ tt”
The blank character string “↓ tt” following the meaningful text Tb1, that is, the second blank character string “↓ tt” in the text Tb appears in the middle of the text Tb (steps S13, S15, S17). This second blank character string “↓ ttt” is determined as the blank character string for formatting in the sixth entry of the blank character information storage unit 55 and stored as the first blank character string “↓ ttt” (FIG. 11 (d)) (steps S19, S21, S23). In this case, the second blank character determination unit 532 uses the first blank character string “↓ ttt” currently processed in the first character in the text Tb stored in the blank character information storage unit 55. As in the case of the blank character string “↓ tt”, it is determined that the blank character string for formatting is related to the structure information “/ patent / abstract / English / text ()” (step S25).

10-4) Meaningful text Tb2
Tb2 following the second blank character string “↓ ttt” in the text Tb is not a blank character string (step S13). In this case, the second blank character determination unit 532 determines that Tb2 is a meaningful text (step S14). The text Tb2 is stored in the text storage unit so as to follow the text Tb1 already stored in the text storage unit.

10-5) Blank character string “↓ tt”
The blank character string “↓ tt” following the meaningful text Tb2, that is, the third blank character string “↓ tt” in the text Tb appears at the end of the text Tb (steps S13, S15, S17).

In this case, the second blank character determination unit 532 determines that the last (third) blank character string “↓ tt” of the text Tb is a formatting blank character string related to the next element, and The character string “↓ tt” is stored in the blank character temporary storage unit 54 in association with the structure information “/ patent / abstract / English” of the corresponding structure (step S18). The document data next to the blank character string “↓ tt” (blank character string at the end of the text Tb) stored in the blank character temporary storage unit 54 is an end tag of the <English> element, and the next element does not exist. The blank character string “↓ tt” is not used as a formatting blank character string.
Thus, the process for the text Tb ends.

(11) </ English>
When the process for the text Tb is completed, the syntax analysis unit 52 performs a structure analysis of the document data subsequent to the text Tb (steps S2 and S3). The document data following the text Tb is a </ English> tag.

  The </ English> tag is an end tag indicating the end of the <English> element indicated by the structure information “/ patent / abstract / English” (steps S4 and S5). In this case, the syntax analysis unit 52 raises the target hierarchy by one level (step S7). As a result, the structure information of the hierarchy to be processed becomes “/ patent / abstract”.

(12) ↓ t
The document data following the </ English sentence> tag is “↓ t”. This “↓ t” is the content (part of) of the <summary> element constituting the current processing target hierarchy, and is “ignorable white space” (steps S4 and S8).

  Therefore, the first blank character determination unit 531 assumes that “↓ t”, which is this “ignorable white space”, is the formatting blank character of the element that appears next, and uses the “↓ t” as the current processing target hierarchy. Corresponding to the structure information “/ patent / abstract” is stored in the blank character temporary storage unit 54 (step S9). FIG. 12B shows the contents of the blank character temporary storage unit 54 at this time. Since the document data next to the blank character string “↓ t” stored in the blank character temporary storage unit 54 is an end tag of the <element> element, the blank character string “↓ t” is used as a formatting blank character string. Not.

(13) </ Summary>
The </ summary> tag is an end tag indicating the end of the <summary> element indicated by the structure information "/ patent / summary" (steps S4 and S5). In this case, the syntax analysis unit 52 raises the target hierarchy by one level (step S7). As a result, the structure information of the hierarchy to be processed becomes “/ patent”.

(14) ↓
</ Summary> The document data following the tag is “↓”. This “↓” is the content of (part of) the <patent> element constituting the current processing target hierarchy, and is “ignorable white space” (steps S4 and S8). In this case, “↓” is stored in the blank character temporary storage 54 in association with the structure information “/ patent” of the current processing target hierarchy (step S9). FIG. 12C shows the contents of the blank character temporary storage unit 54 at this time. Since the document data next to “↓” (ignorable white space) stored in the temporary space character storage unit 54 is an end tag of the <patent> element, and the next element does not exist, the “↓” is a formatting blank. Not used as a string.

(15) </ patent>
Document data following “↓” (ignorable white space) is a </ patent> tag. The </ patent> tag is an end tag indicating the end of the <patent> element in the highest hierarchy indicated by the structure information “/ patent” (steps S4 and S5). In this case, the syntax analysis unit 52 determines the end of the structure information “/ patent” (step S2), and ends the analysis of the XML document.

  At this time, the text “information processing apparatus” is stored in the text storage unit. The text storage unit further stores Ta1, Ta2, and Ta3 for the text Ta, and Tb1 and Tb2 for the text Tb. That is, in the text storage unit, in addition to the text “information processing apparatus”, a part obtained by removing the formatting blank character string from the texts Ta and Tb is stored.

  The index creation unit 56 is based on the text “information processing device” stored in the text storage unit and the portion obtained by removing the formatting blank character string from the texts Ta and Tb, that is, based on the meaningful text. An index (index record) used to search an XML document including the meaningful text is created. Then, the index creating unit 56 stores the created index record in the XML database 42 as a part of the index 422 in association with the XML document original.

  In the example of the XML document described above, it is determined that all blank character strings included in the middle part of the text are formatting blank character strings. If the blank character string included in the intermediate portion of the text is different from the blank character string for formatting of the corresponding structure stored in the blank character information storage portion 55 (steps S19 and S21), it is included in the intermediate portion. It is determined that the blank character string to be displayed is not a formatting blank character string but text (valid text) (step S22). In this case, the determined blank character string is an index creation target by the index creation unit 56.

  In the example of the XML document (second XML document) shown in FIG. 10 (FIG. 9), for simplification of explanation, blank characters (single-byte spaces) included in the English texts Ta1 and Ta2 are valid texts. It is processed as a part. However, half-width spaces included in the texts Ta1 and Ta2 may be processed as blank characters. Even in this case, half-width spaces (blank characters) included in the texts Ta1 and Ta2 are determined as text (valid text) (steps S19, S21, and S22), and thus are the same as the above example.

  In the above embodiment, after the document analysis processing by the syntax analysis unit 52, the index creation processing by the index creation unit 56 is performed. However, the index creation processing by the index creation unit 56 may be performed in parallel with the document analysis processing by the syntax analysis unit 52.

Next, XML document search using the index 422 including the index record created as described above will be briefly described with reference to the flowchart of FIG.
Now, assume that a search request is given from the client terminal 20 to the XML database management system 50 (database server 10) via the network 30 when the user operates the client terminal 20. Here, it is assumed that an XML document search using the word “structure information” as a keyword is requested.

  A search unit (not shown) in the XML database management system 50 is based on a keyword (search keyword) “structure information” designated by a search request from the client terminal 20 and an index 422 (index page included in the index 422). Is searched (step S31). That is, the retrieval unit retrieves all index records including “structure information” as a keyword. The search unit extracts all the XML documents (original text) including “structure information” based on the appearance position information included in the searched index records (step S32). Thereby, for example, the XML document shown in FIG. 6 is inserted even though the line feed character “↓” and the two tab characters “t”) are inserted between the character string “structure information” and the character “information”. Instead, it is extracted as an XML document including “structure information”.

[Modification]
Next, a modification of the above embodiment will be described. A feature of this modification is that schema information (document type definition information) called document type definition (DTD) that defines (describes) the structure of the XML document is attached to the XML document (its substance). In this case, the syntax analysis unit 52 parses the structure of the XML document in advance based on the schema information instead of the XML document (its entity) itself.

FIG. 14 shows an example of DTD (document type definition information) that defines the structure of the XML document of FIG. 6 (FIG. 5). “<! ELEMENT patent (name, abstract)>” on the first line of DTD
Indicates that in the <patent> element, the <name> element and the <subject> element must appear in this order.

Second line of DTD
<! ELEMENT Name (#PCDATA)>
Indicates that character data (#PCDATA) appears in the <name> element.

Similarly, the third line of DTD
<! ELEMENT Abstract (#PCDATA)>
Indicates that character data (#PCDATA) appears in the <summary> element.

  The parsing unit 52 analyzes the DTD, so that the <patent> tag has the highest structure (first level), and the <name> tag and the <summary> tag have the second level structure. It can be recognized in advance. The <name> tag and the <summary> tag are tags of a name element and an abstract element in which character data appears, respectively. In other words, the syntax analysis unit 52 analyzes the DTD attached to the XML document (its substance), thereby generating elements (<name> element and <summary> element) in which character data appears and their structures (depth of the hierarchy). Can be recognized in advance.

  Therefore, if the XML document (entity) shown in FIG. 6 (FIG. 5) read by the document reading unit 51 is analyzed on the premise of the recognized structure, the syntax analysis unit 52 causes character data to appear < Only the text nodes included in the name> element and the <summary> element, that is, only the character string following the <name> tag and the character string following the <summary> tag may be analyzed.

  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. For example, in the above-described embodiment and its modification, the XML document is described as an example of the structured document, but the present invention is not limited thereto. The present invention can be similarly applied to structured documents other than XML documents, such as SGML (Standard Generalized Markup Language) documents.

  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.

The block diagram which shows the hardware constitutions of the client-server system which concerns on one Embodiment of this invention. The block diagram which mainly shows the function structure of the XML database management system shown by FIG. The figure which shows the data structure example of the index stored in an XML database in association with an XML document. The figure which shows a part of flowchart which shows the procedure of the parsing process including the blank character determination for formatting applied in the said embodiment. The figure which shows the remainder of the flowchart which shows the procedure of the parsing process including the blank character determination for formatting applied in the said embodiment. The figure which shows an example of the XML document read by the document reading part. FIG. 6 is a diagram showing the XML document shown in FIG. 5 in a format in which line feed characters and tab characters included in the XML document are replaced with symbols “↓” and “t”. The figure which shows the change of the content of the blank character information storage part accompanying the syntax analysis process of the XML document of FIG. The figure which shows the change of the content of the blank character temporary storage part accompanying the syntax analysis process of the XML document of FIG. The figure which shows the other example of the XML document read by the document reading part. FIG. 10 is a diagram showing the XML document shown in FIG. 9 in a format in which line feed characters and tab characters included in the XML document are replaced with symbols “↓” and “t”. The figure which shows the change of the content of the blank character information storage part accompanying the syntax analysis process of the XML document of FIG. The figure which shows the change of the content of the blank character temporary storage part accompanying the syntax analysis process of the XML document of FIG. The flowchart which shows the procedure of the XML document search using the index applied in the said embodiment. The figure which shows an example of the document type definition information applied in the modification of the said embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 ... Database server, 11 ... Memory, 20 ... Client terminal, 30 ... Network, 40 ... External storage device, 41 ... XML database management program, 42 ... XML database, 50 ... XML database management system, 51 ... Document reading part, 52 ... Parsing unit (ignorable blank character determination means) 53. Blank character determination unit 54. Blank character temporary storage unit 55. Blank character information storage unit 56. Index creation unit 57 57 Document storage processing unit 421 XML document set, 422... Index, 531... First blank character determination unit, 532... Second blank character determination unit, 533.

Claims (8)

A parsing means for analyzing a structure of a structured document to be stored in a structured document database, in which a hierarchical structure is expressed using a tag, and detecting text appearing after the tag;
Blank character information storage for storing a blank character string as a formatting blank character that appears in a structured document analyzed by the syntax analysis means in association with structure information representing a structure associated with the blank character string. Means,
An ignorable blank character judging means for judging whether or not the text detected by the parsing means is an ignorable blank character composed only of white space characters;
When the text is the ignorable whitespace character and the ignorable whitespace character is followed by a start tag, the ignorable whitespace character is the formatting whitespace character related to the structure of the next element including the start tag. First blank character determination means for confirming and storing in the blank character information storage means in association with the structure information representing the structure;
When the text is not the ignorable white space character, the disassembling means for decomposing the text into a white space character string and a text portion not including the white space character string,
At least, determine that the empty string at the beginning of the text that has been decomposed by the decomposition means is operable to determine whether the appearance, is the empty string if the empty string has appeared is the shaping space character A second blank character determination unit that stores the blank character string in the blank character information storage unit in association with structure information representing a structure associated with the blank character string ;
An index used for searching the structured document including the text is created based on the text portion obtained by removing the formatting blank character and the confirmed blank character string from the text, and stored in the structured document database. Indexing means to
A structured document database management system comprising: a document storage processing unit that stores the structured document including the text in the structured document database. The second blank character judging means is a case where a blank character string appears at the head of the text decomposed by the decomposing means, and the character string length of the blank character string is 1 of the structure including the text. Determining that the blank character string is the formatting blank character if it is longer than the character string length of the blank character string stored in the blank character information storage means in association with the structure information at a higher level. The structured document database management system according to claim 1, wherein: The blank character string stored in the blank character information storage means corresponds to flag information indicating whether the blank character string is determined to be the formatting blank character or the formatting blank character candidate. Attached,
The second blank character determining means, when storing the blank character string determined to be the formatting blank character in the blank character information storage means, corresponds to flag information indicating confirmation to the blank character string. In the case where a blank character string appears in the middle part of the text decomposed by the disassembling means, the blank character information is stored in association with the blank character string and the structure information related to the blank character string. It is determined whether it is stored in the means, and at least if it is not stored, the blank character string is associated with the structure information of the structure related to the blank character string and flag information indicating candidates, and the blank character information is stored. accumulate means, if it is accumulated and and associated with the flag information indicating the candidate, according to claim 1, wherein the change in state indicating determine the flag information Structured document database management system. The second blank character judging means is a case where the blank character string appearing in the middle part of the text and the structure information of the structure related to the blank character string are not associated and stored in the blank character information storage means. And the character string length of the blank character string is longer than the character string length of the blank character string stored in the blank character information storage unit in association with the structure information one level higher than the structure including the text. 4. The structure according to claim 3 , wherein the blank character string is stored in the blank character information storage unit in association with the structure information of the structure related to the blank character string and flag information indicating a candidate. Document database management system. When it is determined by the ignorable blank character determination means that it is an ignorable blank character, and when a blank character string appears at the end of the text decomposed by the decomposition means, the corresponding blank character string is formatted. A blank character temporary storage means for temporarily storing the blank character in association with the structure information representing the structure of the element including the blank character string,
The first blank character determination means, when a start tag following the blank character string stored in the blank character temporary storage means is detected by the syntax analysis means, the blank character string as a formatting blank character , structured document database management system of claim 1, wherein the in association with the structural information indicating the structure of the element containing the begin tag accumulated in the space character information Ho蓄 product means. A document reading means for reading the structured document;
The structured document database management system according to claim 1, wherein the syntax analysis unit analyzes the structure of the structured document read by the document reading unit. A document reading means for reading the structured document;
The structured document has document type definition information defining the structure of the structured document,
The parsing means analyzes the document type definition information, thereby recognizing in advance the elements and structures of character data in the structured document to which the document type definition information is attached, and by the document reading means. 2. The structured document database management system according to claim 1, wherein when the structured document is read, the text following the tag of the recognized element in the structured document is analyzed. A database server computer for managing a structured document database in which a structured document in which a hierarchical structure is expressed using tags is stored, and appears in the structured document to be stored in the structured document database; A database server computer having blank character information storage means for storing a blank character string as a formatting blank character in association with structure information representing a structure related to the blank character string,
Parsing means for analyzing the structure of the structured document to be stored in the structured document database and detecting text appearing after the tag;
An ignorable blank character judging means for judging whether or not the text detected by the parsing means is an ignorable blank character composed only of white space characters;
When the text is the ignorable whitespace character and the ignorable whitespace character is followed by a start tag, the ignorable whitespace character is the formatting whitespace character related to the structure of the next element including the start tag. First blank character determination means for confirming and storing in the blank character information storage means in association with the structure information representing the structure;
When the text is not the ignorable white space character, the disassembling means for decomposing the text into a white space character string and a text portion not including the white space character string,
At least, determine that the empty string at the beginning of the text that has been decomposed by the decomposition means is operable to determine whether the appearance, is the empty string if the empty string has appeared is the shaping space character A second blank character determination unit that stores the blank character string in the blank character information storage unit in association with structure information representing a structure associated with the blank character string ;
An index used for searching the structured document including the text is created based on the text portion obtained by removing the formatting blank character and the confirmed blank character string from the text, and stored in the structured document database. Indexing means to
A program for causing the structured document including the text to function as document storage processing means for storing the structured document in the structured document database.

Images