JP4398664B2 - XML data search method and search apparatus, program, and recording medium recording the program - Google Patents

Description

【００３９】
検索例２について
図６は、検索例２を示している。
この検索例２の指定は次の通りである。
例２
/ｒｏｏｔ/ｍｅｍｂｅｒ［ｃｏｍｐａｎｙ＝‘ＳＥＣ’］/ ｎａｍｅ
この指定は、一連のエレメントの指定はルートエレメント（ｒｏｏｔ），メンバーエレメント（ｍｅｍｂｅｒ），ネームエレメント（ｎａｍｅ）であり、検索条件は括弧内の述語である。先頭のエレメントはルートエレメント（ｒｏｏｔ）、最後のエレメントはネームエレメント（ｎａｍｅ）であり、検索起点エレメントは検索条件の直上に位置するメンバーエレメント（ｍｅｍｂｅｒ）である。検索条件はカンパニーエレメント（ｃｏｍｐａｎｙ）がＳＥＣであることが条件となっている。
【００４０】
処理１ａ
この例では、２番目のステップの述語中で最も下のパスにあるエレメントのカンパニーテーブル２０３を検索し、カンパニーテーブル２０３のテキストカラムが‘ＳＥＣ’に合致するレコードを選択し、選択情報として選択レコード番号（１，２）を記憶する。
処理１ｂ
次に、現在のカンパニーテーブル２０３の選択レコード（１，２）のアップグリップカラム（１，４）と直上のメンバーテーブル２０１のダウングリップカラム（１，４）が一致するレコードをジョインし、メンバーテーブル２０１でジョイン対象となったレコード（１，２）を選択する。
このメンバーテーブル２０１は検索起点エレメントに対応するエレメントテーブルであり、処理１ｂはこれで終了する。
【００４１】
処理２
ルートエレメントのルートテーブル２００のレコードのダウングリップカラム（０）と、メンバーテーブル２０１の処理１―ｂで選択された選択レコード（１，２）のアップグリップカラム（０，０）が一致するレコードを結合し、下位のメンバーテーブル２０１でジョイン対象となったレコードを選択する（選択レコード（１，２））。
メンバーテーブル２０１は検索指定の最後のエレメントテーブルではないので、さらに処理を続行する。
すなわち、現在のメンバーテーブル２０１の選択レコード（１，２）のダウングリップカラム（１，４）と、ネームテーブル２０２のレコードのアップグリップカラム（１，４）が一致するレコードを結合し、下位のネームテーブル２０２でジョイン対象となったレコードを選択する（選択レコード（１，２））。
ネームテーブル２０２は最後のエレメントに対応しており、処理２は終了する。
処理２で選択されたネームテーブル２０２の選択レコード（１，２）より、子エレメントを再帰的に復元し、元のＸＭＬデータの断片である次のデータが検索結果となる。
<ｎａｍｅ ｉｄ＝“０９９００”>Ｎａｎｃｙ</ｎａｍｅ>
<ｎａｍｅ ｉｄ＝“１００２７”>Ｊｏｈｎ</ｎａｍｅ>
【００４２】
検索例３について
図７は、検索例３を示している。
この検索例３の指定は次の通りである。
/ｒｏｏｔ/ｍｅｍｂｅｒ
この指定は、一連のエレメントの指定はルートエレメント（ｒｏｏｔ），メンバーエレメント（ｍｅｍｂｅｒ）の２エレメントであり、検索条件は無い。先頭のエレメントはルートエレメント（ｒｏｏｔ）、最後のエレメントはメンバーエレメント（ｍｅｍｂｅｒ）である。
処理１ａ、１ｂ
この例では、検索条件に対応するエレメントテーブルは無いので、処理１ａ，１ｂに相当する処理はない。
【００４３】
処理２
ルートエレメントのルートテーブル２００のレコードのダウングリップカラム（０）と、メンバーテーブル２０１のレコードのアップグリップカラムが一致するレコードを結合し、下位のメンバーテーブル２０１でジョイン対象となったレコードを選択する（選択レコード（１，２，３））。
メンバーテーブル２０１は検索指定の最後のエレメントテーブルなので処理２は終了する。
処理２で選択されたメンバーテーブル２０１の選択レコード（１，２，３）より、子エレメントを再帰的に復元し、元のＸＭＬデータの断片である次のデータが検索結果となる。

【００４４】
この検索例３では、メンバーエレメントの個別要素はすべて選択されることになるが、次の検索例４に示すように、ＸＭＬデータの構造によっては一部が除外される。
【００４５】
図８は、検索例４を示している。
この検索例４のＸＭＬデータは、メンバーエレメント（ｍｅｍｂｅｒ）と並列にエー・エレメント（ａ）があり、このエー・エレメントの子要素としてメンバーエレメント（ｍｅｍｂｅｒ）がつながっている構造である（図８（Ａ），（Ｂ）参照）。
この場合のエレメントテーブルは、図８（Ｃ）に示す通り、エー・エレメントの子要素であるメンバーエレメントの個別データのアップグリップカラムは、エー・テーブルのダウングリップの値４が入れられる。
このようなＸＭＬデータについて、次の検索指定をする。
/ｒｏｏｔ/ｍｅｍｂｅｒ
一連のエレメントの指定はルートエレメント（ｒｏｏｔ），メンバーエレメント（ｍｅｍｂｅｒ）の２エレメントであり、検索条件は無い。先頭のエレメントはルートエレメント（ｒｏｏｔ）、最後のエレメントはメンバーエレメント（ｍｅｍｂｅｒ）である。
【００４６】
処理１ａ、１ｂ
この例では、検索条件に対応するエレメントテーブルは無いので、処理１ａ，１ｂに相当する処理はない。
処理２
ルートエレメントのルートテーブル２００のレコードのダウングリップカラム（０）と、メンバーテーブル２０１のレコードのアップグリップカラムが一致するレコードを結合し、下位のメンバーテーブル２０１でジョイン対象となったレコードを選択する（選択レコード（１，２，３））。
メンバーテーブル２０１は検索指定の最後のエレメントテーブルなので処理２は終了する。
処理２で選択されたメンバーテーブル２０１の選択レコード（１，２，３）より、子エレメントを再帰的に復元し、元のＸＭＬデータの断片である次のデータが検索結果となり、ａの子要素のメンバーエレメントの個別要素はジョイン処理によって除外される。
<ｍｅｍｂｅｒ> … </ｍｅｍｂｅｒ>
<ｍｅｍｂｅｒ> … </ｍｅｍｂｅｒ>
<ｍｅｍｂｅｒ> … </ｍｅｍｂｅｒ>
【００４７】
上記各検索例では、検索指定（ＸＰａｔｈ）の検索条件（述語）内にエレメントが一つしか無いが、ＸＭＬ文書の内容や検索指定によって、検索条件として述語内にエレメントが複数記載される場合がある。
たとえば、
/ｒｏｏｔ［ｍｅｍｂｅｒ/ｃｏｍｐａｎｙ ＝ “ＳＥＣ”］という指定の場合には、検索条件に対応するエレメントはカンパニー（ｃｏｍｐａｎｙ）であり、検索起点エレメントはルート（ｒｏｏｔ）である。
【００４８】
他の実施の形態
上記実施の形態では、検索の指定をＸＰａｔｈに準拠した形で指定するようにしたが、ＸＰａｔｈに準拠する必要はなく、検索条件を含む一連のエレメントを指定できればよい。
【００４９】
【発明の効果】
以上説明したように、本発明によれば、検索条件がある場合には検索条件に合致する個別データのレコードを選択し、この選択した情報を検索条件が付随する検索起点エレメントに対応するエレメントテーブルまで反映させておき、次いで指定した先頭のエレメントから最後のエレメントまで結合処理とと結合結果のレコードの選択を繰り返すことにより、最後のエレメントテーブルについて、検索条件を反映したレコードを選択するようにしたので、
ダウングリップとアップグリップの値を利用して機械的に結合、レコードの選択を繰り返すだけで、取得すべきエレメントの個別データを検索することができる。
【００５０】
取得すべきエレメントの個別データおよびその子孫に当たるエレメントの個別データをＸＭＬデータとして復元することにより、ＸＭＬデータとして利用できる。
【図面の簡単な説明】
【図１】図１（Ａ）は本発明の実施の形態に係るＸＭＬデータの検索装置を含むデータベースの概略構成を示す図、図１（Ｂ）は同図（Ａ）の記憶装置のテーブルに形成されるカラムの種類を示す図である。
【図２】図２（Ａ）はＸＭＬデータの一例を示す図、同図（Ｂ）は同図（Ａ）のデータのリンク関係をツリー構造にして示した図、同図（Ｃ）は同図（Ａ）のデータを格納したテーブルのデータ格納状態を示す図である。
【図３】図３はテーブル間の親子関係情報を記憶するテーブル構成を示す図である。
【図４】図４は検索手順のフローチャートの一例である。
【図５】図２のＸＭＬデータの検索例１の検索手順を模式的に示す説明図である。
【図６】図２のＸＭＬデータの検索例２の検索手順を模式的に示す説明図である。
【図７】図２のＸＭＬデータの検索例３の検索手順を模式的に示す説明図である。
【図８】図２と異なる構造のＸＭＬデータの検索例４の説明図である。
【符号の説明】
１００ ＸＭＬデータ
１０１ ＤＴＤ
１０２ データ本体部
２００，２０１，２０２ テーブル
３００ 格納処理部
４００ 記憶装置
５００ 検索処理部
Ｃ１ 属性カラム
Ｃ２ テキストカラム
Ｃ３ ダウングリップカラム
Ｃ４ アップグリップカラム[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a database of XML data having a hierarchical structure described in XML (Extensible Markup Language), and more particularly to a technique for retrieving XML data stored in a table in a table format.
[0002]
[Prior art]
Conventionally, various methods have been adopted for storing XML data in a database management system and for searching the stored data at high speed.
When storing XML data based on a relational model, there are basically three methods.
[0003]
The first is a method of storing one XML data as one character string data.
In this case, the entire XML data is basically stored as one character string data in a specific column of a table provided for storing XML data.
Also, mainly for the purpose of improving the search performance, a search column is prepared in the same table separately from the column storing the XML data as a character string, and the element or attribute value in the XML data used for the search is set in advance. A method of extracting and storing in the column may be used at the same time.
[0004]
The second method is a method of disassembling each element and attribute of the XML data and storing it in correspondence with each column of a plurality of tables designed and created in advance for the target XML.
In this case, there is a method of extracting and storing only necessary element / attribute values in addition to a method of decomposing and storing all element / attribute values. In the latter case, it may be used in combination with the first method.
[0005]
The third method is to divide the table into an XML data element or attribute information table and a table storing link information representing the parent-child relationship between elements.
In relational databases, in order to speed up the search, index information of the element or attribute value to be searched is created separately, and the target data is searched by referring to the index information first when searching. Techniques are commonly used.
[0006]
[Problems to be solved by the invention]
The problem with the first method is that the entire XML data is handled as one character string information in the database, so it is necessary to create a special index in advance in order to speed up the search. If the information changes, it is necessary to recreate the index.
In addition, it is difficult to perform high-speed range search and aggregation on the value of an element or attribute at a specific position, and the data type of the element or attribute value is limited to the character type. There is.
[0007]
The problem with the second method is that it is necessary to design the database in advance according to the structure of the XML data to be stored. Therefore, if the structure of the XML data changes, it is necessary to start over from the design. This greatly hinders the degree and flexibility.
In general, it is technically difficult to associate XML data having a tree structure with a relational model, and as a result, it is difficult to improve search performance.
[0008]
The problem with the third method is that the link information needs to be traced every time in order to search for information on an element or attribute at a specific position. Depending on the XML data structure and search contents, the link is traced many times. In general, it is difficult to perform a high-speed search process. Further, like the first method, it is difficult to provide data type variations such as numerical types and character types, while being able to flexibly deal with a relatively large number of XML data structures.
Other related technologies include, for example, Japanese Patent Application Laid-Open Nos. 2002-73589, 2001-34618, 2001-318935, 2001-282856, and 2001-195406.
[0009]
The present invention has been made to solve the above-described problems of the prior art, and an object of the present invention is to provide an XML data search method and search apparatus for efficiently searching XML data stored in a table format, and To provide a program and a recording medium on which the program is recorded.
[0010]
[Means for Solving the Problems]
  To achieve the above object, the XML data search method of the present inventionIs
  Create an element table for each element based on the schema information of the XML data to be stored in the database area of the storage device,
In each element table, create a column to store the individual data of the element, a down grip column, and an up grip column,
  The data is stored in each table in order from the root element. Appearance order information is added to the stored data so that all elements have unique values, and the appearance appears in the downgrip column. Enter the order information, and assign the appearance information of the downgrip column of the parent data that has a parent-child relationship with the data to the upgrip column, and store it as a one-line record with the link relationship between the data.XML dataUsing XML data search deviceSearchhow toIt is.
[0011]
The XML data is data having a hierarchical structure described in the well-known XML, and is described with an element (element) as a unit, and a tag indicating the content is attached to the element. By grouping elements, the hierarchical structure of the elements is expressed, and can be expressed as a tree structure (tree structure) model as a whole.
[0012]
Rules such as elements to be used, data contents, element attributes, and parent-child relationships between elements are designed as schemas. For example, DTD is defined as a schema description method, but the description method is not limited to description according to DTD, and various languages such as XML Schema can be used.
[0013]
  The XML data search device includes:Read a search specification for a set of elements including search criteria,
  Search the element table corresponding to the search condition,
  Select the record of individual data that matches the search condition from the individual data stored in the searched element table,
  A record in which the upgrip column of the selected record of the current element table and the downgrip column of the upper element table match are joined using the selection information of the record as a starting point, and the record becomes a join target in the upper element table Select the record that reflects the search condition for the element table of the search origin element by repeating the procedure of selecting to the element table of the search origin element accompanied by the search condition,
  Go to the first element of the search specification,
  Downgrid current element table record or selected recordPukaThe upgrip column of the ram and the lower element table record or selected record match.
Select the record that reflects the search condition for the last element table by repeating the procedure for selecting the records to be joined in the lower element table up to the element table of the last element specified for search.Execute the processIt is characterized byThe
[0014]
  Also,The XML data search device includes:Restore XML data based on the record selected for the last element table specified for searchExecute the processIt is characterized by that.
[0015]
  The XML data retrieval apparatus of the present invention creates an element table for each element based on the schema information of the XML data to be stored in the database area of the storage device, and each element table has a column for storing individual data of the element. Create a down grip column and an up grip column,The data is stored in each table in order from the root element. Appearance order information is added to the stored data so that all elements have unique values, and the appearance appears in the downgrip column. Put the order information, assign the appearance information of the downgrip column of the parent data that has a parent-child relationship with the data to the upgrip column and attach the link relationship between the data,This is a search device for XML data stored as a one-line record.
[0016]
That is, a search specification of a series of elements including a search condition is read, an element table corresponding to the search condition is searched, and individual data records that match the search condition are stored among the individual data stored in the searched element table. Selected,
A record in which the upgrip column of the selected record of the current element table and the downgrip column of the upper element table match are joined using the selection information of the record as a starting point, and the record becomes a join target in the upper element table Select the record that reflects the search condition for the element table of the search origin element by repeating the procedure of selecting to the element table of the search origin element accompanied by the search condition,
Go to the first element of the search specification, and join the records that match the downgrip grip column of the record of the current element table or the selected record and the record of the lower element table or the upgrip column of the selected record. It is provided with processing means for executing a procedure for selecting a record reflecting a search condition for the last element table by repeating the procedure for selecting the record to be combined up to the element table of the last element specified for search. Features.
The processing means restores the XML data based on the record selected for the last element table specified for search.
[0017]
  The present invention also provides a program for causing a computer to execute the XML data search method and a computer-readable recording medium storing the program.
  That is, this program creates an element table for each element based on the schema information of the XML data to be stored in the database area of the storage device,
  In each element table, create a column to store the individual data of the element, a down grip column, and an up grip column,
  The data is stored in each table in order from the root element. Appearance order information is added to the stored data so that all elements have unique values, and the appearance appears in the downgrip column. XML data stored as a one-line record with order information entered and the up-grip column assigned with the appearance order information of the down-grip column of the parent data that has a parent-child relationship with the data, and linked with the data. Search program,
  Read the search specification of a series of elements including the search condition, search the element table corresponding to the search condition, and select the record of individual data that matches the search condition from the individual data stored in the searched element table ,
  A record in which the upgrip column of the selected record of the current element table and the downgrip column of the upper element table match are joined using the selection information of the record as a starting point, and the record becomes a join target in the upper element table Select the record that reflects the search condition for the element table of the search origin element by repeating the procedure for selecting to the element table of the search origin element with the search condition, and go to the first element of the search specification. A record in which the downgrip column in the table record or selected record matches the record in the lower element table or the upgrip column in the lower element table, and the record to be merged in the lower element table The procedure for selecting a procedure for selecting a record that reflects the search for the last element table by repeating until the element table of the last element of the search specificationLet computer runIt is characterized by that.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
The present invention will be described below based on the illustrated embodiments.
FIG. 1A conceptually shows an apparatus configuration of a database including the XML data search apparatus of the present invention.
The database includes a storage device 400 having a database area 401 for storing data, and a storage device 400 for each element defined by DTD 101 (Document Type Definition) in which schema information of XML data 100 is described. A storage processing unit programmed to create a table 200,... In the database area 401 and distribute and store the data of the data main body unit 102 in which the actual tagged data is described in the corresponding tables 200, 201, 202. 300, and a search processing unit 500 that searches data stored in each table 200,... Further, an input operation unit 501 for inputting a search condition to the search processing unit 500 and a display unit 502 for outputting a search result by the search processing unit 500 are provided.
A folder 402 is created in the database area 401, and the tables 200,... Are created in the folder 402 together with the DTD 101.
[0019]
First, the premise data storage will be described, and then the data search will be described.
Although not particularly shown, the storage processing unit 300 and the search processing unit 500 are configured to include a normal computer microprocessor (CPU) and a main storage device such as a RAM, and data is processed according to a program read into the main storage device. The
A main storage device such as a RAM is also used for the storage device 400 in which the tables 200,.
The program is generally stored in an auxiliary storage device such as a hard disk, and is read into the main storage device and executed when processing is executed. The program can be stored in various recording media such as a CD-ROM, a CD-R, a CD-RW, a DVD-ROM, a DVD-R, and a DVD-RW, and can be distributed via a network.
[0020]
XML is a structured data description language that can eliminate the need for description such as DTD as schema information, but the XML data 100 handled by the present invention requires schema information. The schema information is not limited to the DTD 101, and may be information using another schema language such as XML Schema or RELAX NG. In FIG. 1A, the DTD 101 is required only when the folder 402 is first created, and is not required when storing XML data.
[0021]
FIG. 1B illustrates the columns created in the table.
The attribute (attr-) column C1 is one of the columns related to the contents of the individual data, and stores attribute values. When the attribute is omitted, a NULL value is stored.
A text (text) column C2 is one of the columns related to the content of the individual data, and stores the text value of the content of the individual data of the element. This text column is entered only if it has a text value.
An order number is assigned to the down-grip column C3 as appearance order information in the order of appearance of data in the data body 102.
[0022]
The up grip (up-grip) column C4 stores the value of the down grip of the corresponding parent element. When there is no child element having an up grip corresponding to the down grip, it is displayed that there is no data to be a child.
[0023]
To store XML data, first create a table in the folder 402 by specifying the DTD, and store and store the individual data in the corresponding element table column in order from the root element for the XML data according to the DTD. To go. The storage is performed by assigning the appearance order information of the individual data to the down grip column, assigning the appearance order information of the down grip column of the individual data corresponding to the parent of the individual data to the up grip column, and establishing a link relationship between the individual data. Store as a row record.
That is, when the registration start is instructed, the program of the main control unit 300 is executed, the XML data 100 to be stored is analyzed, the elements are acquired, and the element and attribute data are stored in the corresponding tables.
To obtain an element and store it in the table, follow the data in the data body 102 of the XML data from the root, select the corresponding table by referring to the tag name, and if there is an attribute value, store the attribute value in the attribute column. If there is text data, store it in text column C2
[0024]
The values are assigned to the down grip column C3 from the youngest values in the element appearance order. A unique value for all elements. In the up grip column C4, the value of the down grip assigned to the parent element of the element is set. If there are multiple elements with the same value up grip, the order of their appearance is the descending order of the down grip.
When there is no child element having an up grip corresponding to the down grip, this means that there is no child.
Next, it is determined whether or not all elements have been completed. If storage of all elements has not been completed, the next element is acquired and the data storage procedure is repeated.
When storage of all element data is completed, it is determined whether storage of all XML data to be registered has been completed. If storage of all XML data has not ended, the next XML data is analyzed, Repeat the data storage procedure. When the storage of all the XML data is finished, the registration process is finished.
[0025]
2A and 2C show storage examples of XML data and an element table storing the XML data. For the sake of clarity, a tree structure of XML data is shown (see FIG. 2B), and a value of down grip is assigned to individual data of XML data and each individual data (node) of the tree display.
The XML data is stored in the element table as follows.
The route table 200 is selected from the route tag in the first row, and “0” is stored in the down grip column of the route table 200, and “−1” indicating that there is no parent element corresponding to the up grip column C4 is stored.
[0026]
The member table 201 is selected from the member tags in the second row, and “1” is written in the down grip column C3 and “0” is written in the up grip column C4 of the child table 201.
The name table 202 is selected from the name tags on the third line, "2" is entered in the down grip column C3, "1" is entered in the up grip column C4, "09900" is entered in the attribute column C1, and the text column C2 is entered. Enter "Nancy" in
The company table 203 is selected from the company tag in the fourth row, “3” is entered in the down grip column C3, “1” is entered in the up grip column C4, and “SEC” is entered in the text column C2. .
[0027]
The member table 201 is selected from the member tags in the sixth row, and “4” is written in the down grip column C3 and “0” is written in the up grip column C4 in the second row of the child table 201.
The name table 202 is selected from the name tag on the seventh line, "5" is entered in the down grip column C3 of the name table 202, "4" is entered in the up grip column C4, "10027" in the attribute column C1, and text column C2 Enter "John" in
The company table 203 is selected from the company tag in the eighth row, “6” is entered in the down grip column C3, “4” is entered in the up grip column C4, and “SEC” is entered in the text column C2. .
[0028]
The member table 201 is selected from the member tags in the 10th row, and “7” is entered in the downgrip column C3 and “0” is entered in the upgrip column C4 in the 3rd row of the child table 201.
The name table 202 is selected from the name tag on the eleventh line, “8” is entered in the down grip column C3, “7” is entered in the up grip column C4, “12469” is entered in the attribute column C1, and the text column C2 is entered. Enter "Tom" in
The company table 203 is selected from the company tag on the 12th line, “9” is entered in the down grip column C3, “7” is entered in the up grip column C4, and “HAL” is entered in the text column C2. .
When all the data in the data main body 102 is traced, the registration process is terminated.
[0029]
As described above, the XML data stored in the folder stores all the information such as the element contents, attributes, hierarchical structure, and the appearance order of the elements, and the tabular data stored in the folder is stored in the XML data. Can be reproduced.
[0030]
The parent-child relationship between elements is stored in memory separately from the table.
FIG. 3 shows an example in which the parent-child relationship of elements is added.
The illustrated example is an example in which link relationships are stored in a table format. The table 600 includes an element name column 601, a table number column 602 that sequentially assigns a series of numbers to the table name, and a parent for each table. A parent column 603 for storing the table number and a plurality of child columns 604 for storing the number of the table that is a child of each table are provided. This parent-child relationship data is managed in association with each folder 402.
[0031]
In the illustrated example, “root”, “member”, “name”, and “company” are stored in the element name column 601, and 0, 1 are stored in the number column next to the element name column 601. Numbers 2 and 3 are unique.
For the root element in the first row of the table, “−1” (no parent element) is stored in the parent column, and “1” is stored in the child column.
For the member element in the second row of the table 600, “0” is stored in the parent column, and “2” and “3” are stored in the child column.
For the name element in the third row of the table, “1” is stored in the parent column and “−1” (no child element) is stored in the child column.
For the company element in the fourth row of the table, “1” is stored in the parent column and “−1” (no child element) is stored in the child column.
[0032]
Next, the XML data search apparatus of the present invention will be described.
FIG. 4 shows a flowchart of the search processing procedure.
First, a search specification for a series of elements including a search condition is read and the search specification is analyzed (S1).
The search designation is specified by, for example, XPath, for example, from the root element to the element acquired as a result, and the search condition is appropriately specified for each element as necessary.
Next, the element table corresponding to the search condition is searched (S2), and the record of the individual data that matches the search condition is selected from the individual data stored in the searched element table (S3).
[0033]
Using the selection information of the selected record as a starting point, records in which the upgrip column of the selected record of the current element table matches the downgrip column of the upper element table are combined (S5), and the upper element table The procedure for selecting the record to be combined (S6) is repeated up to the element table of the search starting element accompanied by the search condition, thereby selecting the record reflecting the search condition for the element table of the search starting element (S4 to S6). ). Next, the process proceeds to the first element of the search specification (S7), and joins the records in which the down grip grip column of the current element table record or selected record matches the up grip column of the lower element table or selected record. (S9), selecting the record that reflects the search condition for the last element table by repeating the procedure for selecting the record to be combined in the lower element table (S10) and repeating to the element table of the last element specified for search (S8 to S10).
Based on the record selected for the last element table, the XML data is restored and displayed (S11).
[0034]
Next, specific search examples 1 to 3 of XML data will be described with reference to FIGS. XML data is data stored in the table format shown in FIG. 2. In FIGS. 5 to 7, each element table displays only the record number (No) of each row, the column up grip column, and the down grip column. Yes.
The search is specified in a form compliant with XPath.
The designation of XPath is divided into the following two parts.
Specification 1. Specifies the path of a series of elements from the document root to the resulting element.
Specification 2. Write a predicate representing the search condition.
[0035]
In response to this specification, the following processing is performed.
Process 1.
If a predicate (search condition) is specified for each step of XPath, the corresponding element table is searched and a record that matches the condition is selected.
1a. Narrow down the element table in the lowest path in the predicate and create record selection information.
1b. Perform join with the element table of the path directly above, and select the record to be joined in the element table directly above.
This process is repeated up to the element table corresponding to the search starting element accompanied by the search condition, and a record reflecting the search condition is selected for the element table of the search starting element.
For example, in the XPath expression, the search origin element is an element described immediately before (above) a predicate representing a search condition, and is accompanied by a search condition (predicate).
Joining is a process of combining records in which the up grip column of the selected record in the lower element table matches the down grip column of the upper element table.
Process 2.
When a record that reflects the search condition is selected for the element table of the search origin element, join the element table record or the selected record in the step immediately below in order from the document root, and select the join target in the lower element table. The procedure of selecting the record that has become is repeated until the last element table, and the record that reflects the search condition in the last element table is selected.
[0036]
Search example 1
FIG. 5 shows search example 1.
The search specification is as follows.
/ root / member [name / @ id = ’10027’]
That is, the specification of a series of elements is root and member, and the search condition is a predicate in parentheses. The first element is a root element (root), the last element is a member element (member), and the element table of the search origin element is a member element (member) located immediately above the search condition. The search condition is a record whose attribute of the name element (name) is 10027.
[0037]
Process 1a
In this example, the name table 202 of the element in the lowest path in the predicate of the second step is searched, and a record that matches the attribute column of the name table 203, atr-id = '10027' is selected and selected. For example, the selected record number (2) is stored as information.
As the selection information, information for distinguishing from other records may be added to the record storing the corresponding individual data in the table.
Process 1b
Next, a record in which the up grip column (4) of the selected record (2) of the current name table 202 matches the down grip column of the member table 201 immediately above is joined, and the record ( 2) is selected.
This member table 201 is an element table corresponding to the search starting element, and the process 1b is completed here.
[0038]
Process 2
A record in which the down grip column (0) of the record of the route table 200 of the root element matches the up grip column of the selected record 2 selected in the process 1-b of the member table 201 is joined, and the join is performed in the lower member table. The target record is selected (selected record (2)).
The member table 201 corresponds to the last element, and the process 2 ends. .
The child element is recursively restored from the selected record (2) selected in the process 2, and the next data that is a fragment of the original XML data becomes the search result.

[0039]
About search example 2
FIG. 6 shows search example 2.
The designation of this search example 2 is as follows.
Example 2
/ root / member [company = ‘SEC’] / name
In this specification, a series of elements are a root element (root), a member element (member), and a name element (name), and a search condition is a predicate in parentheses. The first element is a root element (root), the last element is a name element (name), and the search origin element is a member element (member) located immediately above the search condition. The search condition is that the company element (company) is SEC.
[0040]
Process 1a
In this example, the company table 203 of the element in the lowest path in the predicate of the second step is searched, the record whose text column of the company table 203 matches “SEC” is selected, and the selected record is selected information. The number (1, 2) is stored.
Process 1b
Next, a record in which the up grip column (1, 4) of the selected record (1, 2) of the current company table 203 matches the down grip column (1, 4) of the member table 201 immediately above is joined, and the member table In 201, the record (1, 2) to be joined is selected.
This member table 201 is an element table corresponding to the search starting element, and the process 1b is completed here.
[0041]
Process 2
A record in which the down grip column (0) of the record of the route table 200 of the root element matches the up grip column (0, 0) of the selected record (1, 2) selected in the process 1-b of the member table 201 is selected. The records to be joined and selected in the lower member table 201 are selected (selected records (1, 2)).
Since the member table 201 is not the last element table specified for search, the processing is further continued.
That is, a record in which the down grip column (1, 4) of the selected record (1, 2) of the current member table 201 matches the up grip column (1, 4) of the record of the name table 202 is joined. A record to be joined in the name table 202 is selected (selected record (1, 2)).
The name table 202 corresponds to the last element, and the process 2 ends.
The child element is recursively restored from the selected record (1, 2) of the name table 202 selected in the process 2, and the next data that is a fragment of the original XML data becomes the search result.
<name id = “09900”> Nancy </ name>
<name id = “10027”> John </ name>
[0042]
About search example 3
FIG. 7 shows search example 3.
The designation of this search example 3 is as follows.
/ root / member
In this designation, a series of elements are designated by two elements, a root element (root) and a member element (member), and there is no search condition. The first element is a root element (root), and the last element is a member element (member).
Process 1a, 1b
In this example, since there is no element table corresponding to the search condition, there is no process corresponding to the processes 1a and 1b.
[0043]
Process 2
A record in which the downgrip column (0) of the record of the route table 200 of the root element matches the upgrip column of the record of the member table 201 is joined, and the record to be joined in the lower member table 201 is selected ( Selection record (1, 2, 3)).
Since the member table 201 is the last element table specified for search, the process 2 ends.
The child elements are recursively restored from the selected record (1, 2, 3) of the member table 201 selected in the process 2, and the next data that is a fragment of the original XML data becomes the search result.

[0044]
In this search example 3, all the individual elements of the member elements are selected, but some are excluded depending on the structure of the XML data, as shown in the following search example 4.
[0045]
FIG. 8 shows search example 4.
The XML data of this search example 4 has a structure in which there is an element A (a) in parallel with a member element (member), and member elements (member) are connected as child elements of the element A (FIG. 8 ( (See A) and (B)).
In the element table in this case, as shown in FIG. 8C, the up grip column of the individual data of the member element which is a child element of the A element is set with the value 4 of the down grip of the A table.
The following search is designated for such XML data.
/ root / member
A series of elements are designated by two elements, a root element (root) and a member element (member), and there is no search condition. The first element is a root element (root), and the last element is a member element (member).
[0046]
Process 1a, 1b
In this example, since there is no element table corresponding to the search condition, there is no process corresponding to the processes 1a and 1b.
Process 2
A record in which the downgrip column (0) of the record of the route table 200 of the root element matches the upgrip column of the record of the member table 201 is joined, and the record to be joined in the lower member table 201 is selected ( Selection record (1, 2, 3)).
Since the member table 201 is the last element table specified for search, the process 2 ends.
The child element is recursively restored from the selected record (1, 2, 3) of the member table 201 selected in the process 2, and the next data that is a fragment of the original XML data becomes the search result, and the child element of a Individual elements of the member elements are excluded by the join process.
<member>… </ member>
<member>… </ member>
<member>… </ member>
[0047]
In each of the above search examples, there is only one element in the search condition (predicate) of the search specification (XPath). However, depending on the content of the XML document and the search specification, a plurality of elements may be described in the predicate as the search condition. is there.
For example,
When / root [member / company = “SEC”] is specified, the element corresponding to the search condition is a company, and the search start element is a root.
[0048]
Other embodiments
In the above embodiment, the search specification is specified in a form that conforms to XPath. However, it is not necessary to conform to XPath, and it is sufficient that a series of elements including search conditions can be specified.
[0049]
【The invention's effect】
As described above, according to the present invention, when there is a search condition, an individual data record that matches the search condition is selected, and this selected information is used as an element table corresponding to the search origin element accompanied by the search condition. Then, the record that reflects the search condition is selected for the last element table by repeating the join process and the join result record selection from the specified first element to the last element. So
By simply combining and selecting records using the down grip and up grip values, individual data of elements to be acquired can be searched.
[0050]
The individual data of the element to be acquired and the individual data of the element corresponding to the descendant are restored as XML data, and can be used as XML data.
[Brief description of the drawings]
FIG. 1A is a diagram showing a schematic configuration of a database including an XML data search device according to an embodiment of the present invention, and FIG. 1B is a table of a storage device in FIG. It is a figure which shows the kind of column formed.
2A is a diagram showing an example of XML data, FIG. 2B is a diagram showing a data link relation of FIG. 2A in a tree structure, and FIG. It is a figure which shows the data storage state of the table which stored the data of FIG.
FIG. 3 is a diagram showing a table configuration for storing parent-child relationship information between tables.
FIG. 4 is an example of a flowchart of a search procedure.
FIG. 5 is an explanatory diagram schematically showing a search procedure of XML data search example 1 in FIG. 2;
6 is an explanatory diagram schematically showing a search procedure of XML data search example 2 in FIG. 2; FIG.
7 is an explanatory view schematically showing a search procedure of XML data search example 3 in FIG. 2; FIG.
8 is an explanatory diagram of search example 4 for XML data having a structure different from that in FIG. 2; FIG.
[Explanation of symbols]
100 XML data
101 DTD
102 Data body
200, 201, 202 tables
300 Storage processing unit
400 storage device
500 Search processing part
C1 attribute column
C2 text column
C3 down grip column
C4 Upgrip column

Claims (6)

Create an element table for each element based on the schema information of the XML data to be stored in the database area of the storage device,
In each element table, create a column to store the individual data of the element, a down grip column, and an up grip column,
The data is stored in each table in order from the root element. Appearance order information is added to the stored data so that all elements have unique values, and the appearance appears in the downgrip column. XML data stored as a one-line record with order information entered and the up-grip column assigned with the appearance order information of the down-grip column of the parent data that has a parent-child relationship with the data, and linked with the data. Is a method for searching for data using a search device for XML data ,
The XML data search apparatus reads a search specification of a series of elements including a search condition,
Search the element table corresponding to the search condition,
Select the record of individual data that matches the search condition from the individual data stored in the searched element table,
A record in which the upgrip column of the selected record of the current element table and the downgrip column of the upper element table match are joined using the selection information of the record as a starting point, and the record becomes a join target in the upper element table Select the record that reflects the search condition for the element table of the search origin element by repeating the procedure of selecting to the element table of the search origin element accompanied by the search condition,
Go to the first element of the search specification,
To merge records that match the downgrip column of the current element table record or selected record with the upgrip column of the lower element table or selected record, and select the record to be merged in the lower element table The XML data search method is characterized by executing a process of selecting a record reflecting the search condition for the last element table by repeating the above to the element table of the last element designated for search. The XML data search method according to claim 1, wherein the XML data search apparatus executes processing for restoring XML data based on a record selected for a last element table specified for search. Create an element table for each element based on the schema information of the XML data to be stored in the database area of the storage device,
In each element table, create a column to store the individual data of the element, a down grip column, and an up grip column,
The data is stored in each table in order from the root element. Appearance order information is added to the stored data so that all elements have unique values, and the appearance appears in the downgrip column. XML data stored as a one-line record with order information entered and the up-grip column assigned with the appearance order information of the down-grip column of the parent data that has a parent-child relationship with the data, and linked with the data. Search device,
Read a search specification for a set of elements including search criteria,
Search the element table corresponding to the search condition,
Select the record of individual data that matches the search condition from the individual data stored in the searched element table,
A record in which the upgrip column of the selected record of the current element table and the downgrip column of the upper element table match are joined using the selection information of the record as a starting point, and the record becomes a join target in the upper element table Select the record that reflects the search condition for the element table of the search origin element by repeating the procedure of selecting to the element table of the search origin element accompanied by the search condition,
Go to the first element of the search specification, and combine the downgrip column of the current element table record or selected record with the record that matches the lower element table record or the upgrip column of the selected record, and combine them in the lower element table It is provided with processing means for executing a procedure for selecting a record reflecting a search condition for the last element table by repeating the procedure for selecting a target record up to the element table of the last element specified for search. XML data retrieval device.   4. The XML data search apparatus according to claim 3, wherein the processing means restores the XML data based on the record selected for the last element table specified for search. Create an element table for each element based on the schema information of the XML data to be stored in the database area of the storage device,
In each element table, create a column to store the individual data of the element, a down grip column, and an up grip column,
The data is stored in each table in order from the root element. Appearance order information is added to the stored data so that all elements have unique values, and the appearance appears in the downgrip column. XML data stored as a one-line record with order information entered and the up-grip column assigned with the appearance order information of the down-grip column of the parent data that has a parent-child relationship with the data, and linked with the data. Search program,
Read the search specification of a series of elements including the search condition, search the element table corresponding to the search condition, and select the record of the individual data that matches the search condition from the individual data stored in the searched element table ,
A record in which the upgrip column of the selected record of the current element table and the downgrip column of the upper element table match are joined, and the record to be joined in the upper element table is based on the selection information of the record. Select the record that reflects the search condition for the element table of the search origin element by repeating the procedure for selecting to the element table of the search origin element with the search condition, and go to the first element of the search specification. combine records up grip columns of a record or selected records down grip columns and the lower element table table records or selected records match, was the binding partner at a lower element tables record The procedure for selecting, searching specification of XML data retrieval program, characterized in that to execute the last of the last element table by repeating until the element table of the elements of the procedure for selecting a record search conditions reflect the computer.   A computer-readable recording medium on which the program according to claim 5 is recorded.

Images