WO2014061846A1

WO2014061846A1 - Method for generating index for processing mixed query, method for processing mixed query, and recording medium for recording index material structure

Info

Publication number: WO2014061846A1
Application number: PCT/KR2012/008713
Authority: WO
Inventors: 김신유; 한혁; 진성일; 김재광; 백성인
Original assignee: 주식회사 리얼타임테크
Priority date: 2012-10-17
Filing date: 2012-10-23
Publication date: 2014-04-24
Also published as: KR20140049326A; KR101440475B1

Abstract

The present invention relates to a method for generating an index for processing a mixed query, a method for processing a mixed query, and a recording medium for recording an index material structure. The method for generating the index for processing the mixed query, according to the present invention, comprises the steps of: determining reference column information and expanded column information from factors included in an index generation command, when the index generation command with respect to a specific table in a database is detected; confirming column types of each of columns corresponding to the reference column information and the expanded column by using a table schema of the specific table; and generating and saving to a database management system (DBMS) catalog index key meta information including all of the column types that are confirmed in the previous step.

Description

【Specification】

[Name of invention]

Recording medium recording index generation method, mixed query processing method, and index data structure for complex query processing

Technical Field

The present invention relates to an index generating method for a mixed query processing, a mixed query processing method, and a recording medium recording an index data structure. More specifically, the present invention relates to an index generation method for a mixed query processing including index key data and non-index data. It relates to laws, methods of handling complex queries, and recording media that record these index data structures.

Background Art

In order to retrieve the data stored in the database, a variety of indexing techniques are used.

For example, each table in the database may contain at least one record, and each record may contain at least one column of data. In order to retrieve data stored in the table, the entire table record may be retrieved. Since comparing and searching sequentially is inefficient, create an index for a specific column in advance, and try to search quickly using the index if necessary.

An example of an index structure applied to such an indexing algorithm is the B + tree structure.

The B + Tree consists of two parts, an index part and a leaf sequential set part that stores key values. The index part contains information about a path for quick and direct access to the key on the leaf.

However, only the index column can be specified in the index search condition when creating a conventional B + tree. When the condition is evaluated by a query condition including both the index column and the non-index column, the index column is applied to the index column. Condition evaluation is performed first, and condition evaluation is performed by accessing a table record to obtain column data to perform condition evaluation on non-index columns. In this case, disk 10 may be generated to obtain data close to the record, resulting in performance overhead for condition evaluation.

That is, the conventional B + Tree stores only index column data in the index, so that the index column can be quickly searched. For example, a query having both an index column and a non-index column as a condition, that is, a mixed query In case of input, the B + Tree is used to search only the index column, and the non-index column data for each searched index column must be read to the actual table record. will be.

[Detailed Description of the Invention]

SUMMARY OF THE INVENTION The present invention has been made to solve the above disadvantages, and its object is to develop an index structure capable of processing a mixed query including both index search conditions and non-index search conditions.

Another object of the present invention is to provide a method for processing a mixed query including both an index search condition and a non-index search condition.

It is still another object of the present invention to provide a recording medium which records an index data structure for processing the above-mentioned mixed query. In order to achieve the above object, according to the present invention, an index generation method for a mixed query process detects an index generation command for a specific table in a database, and converts argument factor based column information and extended column information included in the index generation command. Cutting; Identifying each column type of a column subjected to reference column information and a column subjected to extension column information by using a table schema for the specific table; And generating index key meta information including all column types identified in the above step and storing the index key meta information in a DBMS catalog.

Here, at least one column corresponding to the reference column information is determined as an index key column, at least one column referred to as the extended column information is determined as a non-index column, and the index key meta information is determined. And distinguishing the information on the existing index key column and the information on the non-index column.

Here, the index key meta information generated in step (c) includes the number of index key columns, the field meta list for the index key column, the number of non-index columns, and the non-index column meta list. The field meta list includes field meta information for each of the determined index key columns, and the field meta list for the non-index column includes field meta information for each of the determined non-index columns. The field meta information includes a corresponding column type, a position of a corresponding column in a record, a position of a corresponding column in a node entry which is the basis of data retrieval, and a maximum size of the corresponding column data. Here, after the step (c), using the index key meta information to generate a node page for storing at least one node entry that is the basis of data search to perform the initialization of the generated node page Characterized in that it further comprises.

Extracting index key data corresponding to at least one index key column and non-index data corresponding to at least one non-index column from a specific record of a table by referring to the index key meta information; And generating a node entry including all of the extracted at least one index key data and at least one non-index data and storing the extracted node entry in the node page.

The method may further include generating a node item including the location and size of the generated node entry and storing the node item in the node page.

Here, in the step (f), each node item to be treated for each node entry is arranged on the basis of the data to be indexed in the index key column included in the corresponding node entry and stored on the node page. .

In addition, to achieve the above object, a mixed query processing method according to the present invention comprises the steps of: receiving a complex query for data retrieval; Selecting an index most closely related to all columns included in the mixed query; Using the index key meta information of the selected index to identify a first node entry and a last node entry including the same index key data as the index key data included in the mixed query among the node entries included in the selected index; ; A node ent including the same non-index data as the non-index data included in the mixed query while moving from the identified first node entry to the last node entry. Extracting the li;

In addition, to achieve the above object, a mixed query processing method according to the present invention comprises the steps of: receiving a mixed query for data retrieval; Selecting an index most closely related to all columns included in the mixed query; Identifying the node entry included in the selected index using index key meta information of the selected index, identifying a first node entry and a last node entry including the same index key data as the index key data included in the mixed query; ; Extracting non-index data for each non-index column included in the mixed query from each node entry while moving from the identified first node entry to the last node entry.

A computer readable recording medium having recorded thereon an index data structure including at least one index node for performing a data search according to the present invention for achieving the above object, wherein each index node is the basis of data search. At least one node entry is included, and each node entry has a plurality of storage areas each storing data extracted from different columns of any one record in a database table.

Here, at least one of a plurality of storage areas provided in each node entry stores index key data for an index key column of the table, and the non-index column of the table is stored in the remaining storage areas of the plurality of storage areas. Non-indexed data is stored in the characterized in that is stored.

Here, each storage area for storing the data extracted from each column may include a data storage area for storing real data and a data information storage area for storing the location and size of the data storage area.

Here, each index node is characterized in that it comprises at least one node item of a fixed size in which the position and size of each node entry are stored. Here, the total size of each node entry varies according to the number of columns to be stored and the size of actual data.

[Brief Description of Drawings]

1 is a structure of a node page generated by the index generation method according to an embodiment of the present invention,

2 is a view briefly showing a relationship between nodes in a conventional B + Tree;

3 is a structure of a node entry generated by an index generation method according to an embodiment of the present invention,

4 is a detailed structure of a data storage area included in the node entry of FIG. 3;

5 is a detailed structure of a data information storage area included in the node entry of FIG. 3;

6 is a structure of a node item generated by an index generation method according to an embodiment of the present invention '

7 is another example of a node item and a node page generated by an index generation method according to an embodiment of the present invention.

8 is an example of an index key meta data structure;

9 is an example of a table schema for a table generating an index by the index generating method according to an embodiment of the present invention. 10 is an example of index key meta information generated by an index generation method according to an embodiment of the present invention;

11 is an example of data input to the table of FIG. 9;

12 is an example of an index generated by an index generation method according to an embodiment of the present invention;

13 is a diagram illustrating an example of a plurality of node entries having the same index key data.

EXAMPLE

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

First, a node page for each index node of an index for mixed query processing according to an embodiment of the present invention is illustrated in FIG. 1.

In the description of the present embodiment, the index for the mixed query processing uses an example of a modification of the conventional B + Tree.

B + Tree complements the problem of B-Tree and is a tree structure that minimizes the search between nodes in the tree structure for sequential data processing. The key values in a node are kept in ascending order, and the leaf nodes are linked to each other in the form of a linked list, and all the actual data exists only in the leaf node.

An example in which the conventional B + Tree is formed is illustrated in FIG. 2, where one node page is stored in each node (the circular shape of FIG. 2 corresponds to each node), and each node page has a plurality of nodes. The index structure according to an embodiment of the present invention and the conventional B + Tree index structure are the same in that node entries may be included. In this case, the node entry may include at least one index key data as a basis of data search using an index.

The index structure according to an embodiment of the present invention is different from the conventional B + Tree structure in that a node item is further configured as described later in the specific configuration of the node entry. Hereinafter, an index having an index structure according to an embodiment of the present invention will be referred to as 'B + Tree extended index'.

Hereinafter, the difference between the B + Tree extended index structure and the conventional B + Tree structure according to an embodiment of the present invention will be described.

Each node page according to an embodiment of the present invention is as shown in FIG. 1 as described above, wherein the node entry stores data of an index key column, that is, data of an index key data and a non-index column, that is, non-index data. It differs from the conventional in that it has at least one storage area.

That is, each node entry has a plurality of storage areas each storing data extracted from different columns of any one record in the database table. At least one of the plurality of storage areas included in each node entry The index key data corresponding to the index key column of the table is stored, and the non-index data corresponding to the non-index column of the table is stored in the remaining storage areas of the plurality of storage areas.

In this case, the index key column refers to a column that is the basis of automatic sorting when creating a B + Tree extended index of a column included in a table, and a non-index column is not the basis of an automatic sorting when generating a B + Tree extended index. It does not mean a column.

For example, B + Tree extended index each time a new record is added to the table Is updated while maintaining the sort, where the sorting criteria within the B + Tree extended index corresponds to the index key column. According to this embodiment

Automatic sorting of B + Tree extension indexes is done using node items rather than node entries, which will be described later.

Creating indexes here includes creating index structures and adding index data.

As an example of the node entry configuration, when the node entry is divided into two storage areas, the first node entry may include one index key data and one non-index data extracted from the first record. In the case where the node entry is divided into three storage areas, the first node entry may include one index key data extracted from the first record and two non-index data.

In particular, each storage area for storing data extracted from each column may include a data storage area for storing real data and a data information storage area for storing the location and size of the data storage area.

A specific example of such a node entry will be described with reference to FIG. 3.

As shown in the figure, the storage area of one entry can be largely divided into a data information storage area and a data storage area.

First, the data storage area will be described first. The data storage area is an area in which real data of each column is stored, and may have a structure as shown in FIG. 4A.

In the example shown in FIG. 4, index key data is stored in a first storage area, and non-index data is stored in a second storage area and a third storage area, respectively.

4 is only an example, and index key data and non-index data may be stored in different storage areas. In particular, the index key data may be stored in the second storage area and the third storage area. However, considering the sorting and searching functions of the B + Tree extended index, it is desirable to store the index key data in the first storage area.

Next, the data information storage area will be described. The data information storage area is an area for storing information on the data storage area described above, and may have a structure as shown in FIG. 5.

Referring to FIG. 5A, index key data information is stored in a first storage area of a data information storage area, and non-index data information is stored in a second storage area and a third storage area, respectively.

This configuration is just one example, and each storage area may store index key data information or non-index data information. However, considering the sorting and searching capabilities of the B + Tree extended index, the first storage area is indexed. Preferably, key data information is stored.

The order of each column data information (ie, index key data information and non-index data information) stored in the data information storage area is meta information that manages key configuration information of the B + Tree extended index for mixed query processing. The order of the index key columns and non-index columns stored in the key meta information should be identical. In particular, the index key data information should be stored first. In this case, you can store non-index data information after the location where the index key data information is stored. The number of index key data information and non-index data information stored in the data information storage area must match the number of index key data and non-index data stored in the data storage area.

The size of each field in which index key data information and non-index data information are stored must be the same in order to ensure the validity of access to the corresponding information, and each field stores location information of the data storage area and data size information.

That is, each data information storage area may store information about the position and size of each column data. For example, in FIG. 5 (a), the first index key data information may include information about the first index key data of FIG. The position (offsetl) and its size (lengthl) are stored, and if the position (offsets and the size (leghth2)) of the second non-index data of FIG. 4 is stored in the second non-index data information, the third non-index data is stored. The location information stores the position offset3 and the size legacy3 for the third non-index data of FIG. 4.

The storage order for each column of the data information storage area and the storage order for each column of the data storage area should be the same.

If the maximum size of the index key data or non-index data stored in the index key meta information is smaller than the size of each storage area for storing information, the data may be stored directly in the data information storage area for index capacity optimization.

For example, FIG. 5B illustrates an example in which non-index data is directly stored instead of location and size information of non-index data stored in the data storage area in the third storage area of the data information storage area.

On the other hand, a node item in a B + Tree extension index is a data structure that contains node entry information.

Since the total size of each node entry is variable due to the node entry and structure, if the node entries are arranged without a separate management structure for the node entries, the structure of the node entry becomes more complicated and the computational speed may decrease.

Therefore, the node item data structure is defined to store the node entry information stored in the node page in the node item so that the node items can be used.

These node items store two pieces of information.

One piece of information is lower node information. In the conventional B + Tree index structure, the lower node information is stored in the index key data storage area of the parent node, but the B + Tree extension for mixed query processing according to an embodiment of the present invention is performed. The index stores subnode information in the node item structure.

Other information stored in the node entry is information about the node entry, for example, position and size information of the node entry. An example of such a node item structure is shown in FIG. 6.

On the other hand, when creating an index in which only one index key data exists and another non-index data does not exist, the index entry data may be stored in the node item itself because unnecessary capacity overhead is generated by using the node entry structure. . In this case, the node item structure is shown in FIG. 7 (a), and the node page including the node item is shown in FIG. 7 (b).

In this case, however, the condition for storing the index key data in the node item must be clearly stated so that, for example, the maximum size of the index key data Index key data can be stored in the node item only when the size of the region for storing the node entry information is equal to or smaller than the size of the region for storing the node entry information (for example, the size of the region for storing the node entry location and the size).

Hereinafter, for a mixed query processing according to an embodiment of the present invention

How to create a B + Tree extension index.

In order to store data in the generated index or to perform a search using the generated index, meta information representing the structure of the data stored in the index, that is, index key meta information is required.

Therefore, hereinafter, the structure of the index key meta information for the B + Tree extended index for the mixed query processing according to the embodiment of the present invention will be described first. When generating a B + Tree extended index for processing a mixed query, reference column information and extended column information may be input by a user. For example, the reference column information may correspond to index key column information and extended column information. Can show non-index column information. If index key meta information using the input information is generated and stored in DBMS system catalog information, it can be referred to when index data loading and mixed query evaluation (data search) are performed.

The data structure of the index key meta information for the mixed query stored in the DBMS catalog consists of the field meta data structure that stores the column information of the node entry and the key meta data structure that manages the field information data structure.

The field meta data structure is the column type of the index key column or non-index column, the column index where the column is stored in the table record, the maximum size of that column, and the node entry where the index key column data or non-index column data is stored. It includes location information of the data information storage area.

The key meta data structure is a data structure for storing and managing the field meta data structure. The meta information that is to be managed is stored as the number of index key columns and non-index columns composed of field meta information. Include field meta data structures for non-index columns. Furthermore, the key meta data structure may include index type information that specifies the type of node page.

An example of the data structure of such index key meta information is shown in FIG. 8, and a specific example of the index key meta information generated according to this data structure will be described later.

After the index key meta information is generated, the node page may be initialized by creating the node page using the index key meta information before the actual index data is stored.

The initialization process of the node page for each index node may include the process of enlarging the header of the node page. Since the creation and initialization of such a node page is a well-known technique, a detailed description thereof will be omitted. .

Hereinafter, the process of loading index data in the generated index will be described.

Extracts index key data corresponding to at least one index key column and non-index data for at least one non-index column from a specific record of a table by referring to index key meta information stored in a DBMS catalog when a predetermined event occurs. .

In this case, a predetermined event indicates that a new record is registered in the table. It may correspond to a vent, or it may correspond to a command input by a user to extract index data for all records stored in a specific table and load the index data into a pre-generated index.

In addition, a node entry including both the extracted at least one index key data and at least one non-index data is generated and stored in the node page. Information stored in the node entry has been described above. Subsequently, a node item including the location and size of the node entry created in the previous step is generated and stored in the node page.

Hereinafter, specific examples of generating a node page by generating index key meta information on a specific table will be described with reference to FIGS. 9 to 12.

First, it is assumed that the table schema according to the present embodiment is as shown in FIG. It is assumed that the table name of FIG. 9 is 'Student'.

That is, there are six columns in total, each of which has an id, name, address, telephone, sex, age, and the column type and the maximum string for each column are shown in FIG. 9.

Here, the maximum string when the column type is 'int' is represented by '-1', which means that the integer type means that the size varies from system to system. For example, in the case of int type, 4 bytes are system. Can be assigned by

When a user wants to create a B + Tree extended index by selecting 'name' column as an index key column and selecting 'telephone' column and 'age' column as non-index columns for a table with such a table schema. For example, enter the following command (query).

create BPLUSTREE index idx Student on table Student (name) extend telephone, age;

'BPLUSTREE index' means to create a B + Tree extended index, 'idxStudent' is the name of the index to create, 'Student' is the name of the table to be indexed, and '(name)' is 'name' 'Means to select a column as a reference column, that is, an index key column, and' extend telephone, age 'means to select' telephone 'and' age 'columns as non-index columns.

The index key meta information is generated according to the query command. An example of the generated index key meta information is shown in FIG.

The meaning of each line in FIG. 10 will be described with reference to FIG. 8.

'type' is the index type, 'KeyCount' is the number of index key columns, 'KeyFieldDesc' is the field metalist for the index key column, ' _C0 lType' is the column type, and 'colldx' is the column position in the record, 'colPos' Is the position of the column in the node entry, 'cardinality' is the maximum size of the column data, 'ExtCount' is the number of non-index columns, 'ExtFiddDesc' is the field metalist for the first non-index column, and 'ExtFieldDesc2' is the second non-index Field meta list for the column.

Here, the index type indicates whether the node page has a form including both node items and node entries as shown in FIG. 1 or whether the node page has a form including only node items as shown in FIG.

These index types include the number of index key columns selected by the user, the number of non-index columns and their maximum size (maximum string size for character types or space size for storing numeric types). The amount of space where node entry information is stored in the node item (the value is preset by the system. Can be selected).

It is assumed that predetermined data is inserted into the table as shown in FIG. 11 while an index having the index key meta information as shown in FIG. 10 is generated.

11 shows an example in which a total of six records are added to the table. In this case, a node page as shown in FIG. 12 is generated.

Referring to FIG. 12, it can be seen that node entries are stored from the bottom in the record input order of FIG. 11, and node items generated for each node entry are generated with the position and size information of the node entry stored. At this time, the node items are sorted based on the 'name' column data which is the index key data.

In particular, it can be seen that the data information storage area of the node entry has a structure as shown in FIG. That is, the second non-index data 'age' column data is not stored in the data entry area (see FIGS. 3 and 4) of the node entry, but in the data information storage area of the node entry. This is based on the assumption that the size of the data entry storage area of the node entry is equal to or larger than the size of 'int', a type of the 'age' column.

As shown in FIG. 12, not only the 'name' column data, which is index key data, exists in the node entry of the node page, but also 'telephone' column data and 'age' column data, which are non-index data, are also present. Therefore, when a mixed query is input, a quick result can be derived by referring to an index as shown in FIG. 12.

A process of processing a mixed query with reference to the index of FIG. 12 is as follows.

As an example, suppose the following query statement is entered.

SELECT * FROM Student WHERE name = 'Yoo Jae-seok' AND age = 41; In this case, the where statement contains two conditions via the 'and' operator.

The database system first selects the index that most closely matches these two conditions. For example, by checking the index key meta information, the database system can select an index that includes both 'name' and 'age'. As shown in FIG. 10, the index 'idxStudent' includes the 'age' column as a non-index column, although the index key column is set only for the 'name' column, so the database system queries the 'idxStudent' index. You can also choose the index that most closely matches the condition.

Eu database system must first obtain the address and the size of the node entry that contains the 'Yoo' After you find the information quickly aligned node item 'Yoo' using in the _'name' column index key columns of data

Referring to the node item of FIG. 12, it can be seen that the location of the node entry including 'Yoo Jae-seok' is '16330' and its size is '27'.

The database system then moves to the corresponding node entry using the acquired node entry position and size information to read the stored information.

The database system analyzes the condition for non-index data by referring to the information stored in the corresponding node entry. That is, the database system reads the 'age' column data from the node entry and checks whether it corresponds to '41'. This allows the database system to enter a B + Tree extended index when a mixed query containing both index key data and non-index data is entered. All conditional analysis can be performed in to extract node entries that satisfy all of these conditions.

That is, conventionally, since the non-index data is not stored in the node entry, in order to analyze the condition on the non-index data, it was necessary to access the table again and read the necessary data, but according to an embodiment of the present invention, According to FIG. 12, since the non-index data is also stored in the node entry together with the index key data, the condition analysis on the non-index data can be performed in the index.

In the above-described example, the description is made on the assumption that the index key data has not been increased. However, the index key data stored in each node entry may be increased or decreased.

In this case, the database system checks the first node entry and the last node entry that contain the same index key data as the index key data included in the node entry incremental query included in the index, and the last node entry from that first node entry. You can also extract node entries that contain non-index data that is identical to the non-index data included in the mixed query as you move to the entry.

For example, when three node entries having non-index data called 'Yoo Jae-seok' are found as shown in FIG. 13, the database system sequentially moves each node entry, and the node entry having 'age' of 41 is displayed. You can determine if it exists.

As another example, suppose the following Chur statement is entered.

SELECT age FROM Student WHERE name = '유재석';

In this case, the where statement contains only one condition, but the data to be found is the age column data.

Therefore, even in this case, the database system may select an index including both 'name' and 'age' column data, that is, the above-mentioned 'idxStudent' index as an index for searching.

At this time, the process of searching for a node entry with name = 'Yoo Jae-seok' is as shown in the previous example.

After finding a node entry with name = 'Yoo Jae-seok', the database system can extract 'age' column data included in the node entry. That is, even in this case, the database system can process in the index without having to access the table to extract 'age' column data.

On the other hand, the process of performing each of the above-described embodiments can be performed by a program stored in a predetermined recording medium (for example, computer readable).

In the above-described embodiment, one example is to process a mixed query by generating a B + Tree extended index having a structure similar to that of B + Tree, but the present invention is not necessarily limited to the B + Tree extended index. In other words, if both the index entry data and the non-index data are included in the node entry that is the basis of the search, they are within the scope of the present invention.

In addition, the present invention is not limited to the above-described specific embodiments, but may be modified and modified in various ways without departing from the spirit of the present invention. If such changes and modifications fall within the scope of the appended claims, it will be apparent that they are included in the present invention. Industrial availability

As described above, according to the present invention, an index structure for processing a mixed query can be developed and a high performance search result can be expected for a mixed query including both an index search condition and a non-index search condition.

In particular, when utilizing the present invention, when evaluating a mixed query of a database system operating in a mobile environment where resources are limited, unnecessary disks can be evaluated using the data stored in the index without directly accessing the table records. By reducing the performance overhead due to 10, we can expect high performance processing for mixed query evaluation.

Claims

[Claim]

[Claim 1]

(a) determining a reference column information and extended column information among arguments included in the index generation command when detecting an index generation command for a specific table in the database;

(b) identifying each column type for a column treated with reference column information and a column treated with extended column information using the table schema for the specific table;

(c) generating index key meta information including all the column types identified in step (b) and storing the index key meta information in a DBMS catalog.

[Claim 2]

The method of claim 1,

In the step (c),

The index key column is determined for at least one column referred to in the reference column information, the non-index column is determined in at least one column referred to in the extended column information, and the index key column is included in the index key meta information. And distinguishing and including information about the non-index column.

[Claim 3]

The method of claim 2,

The index key meta information generated in step (c) includes the number of index key columns, the field meta list for the index key column, the number of non index columns, and the non index column meta list,

The field meta list for the index key column includes field meta information for each of the determined index key columns.

The field meta list for the non-index column includes field meta information for each of the determined non-index columns.

Each field meta-information includes the corresponding column type, the position of the corresponding column in the record, the position of the corresponding column in the node entry which is the basis of data retrieval, and the maximum size of the corresponding column data. Phosphorus generation method.

[Claim 4]

The method of claim 2,

After step (c),

Generating a node page for storing at least one node entry that is the basis of data search by using the index key meta information and performing the initialization of the generated node page further comprising a mixed query How to index for processing.

[Claim 5]

The method of claim 4,

(d) extracting index key data for at least one index key column and non-index data for at least one non-index column from a specific record of a table with reference to the index key meta information;

(e) the extracted at least one index key data and at least one colorimetric And generating and storing a node entry including all of the data in the node page.

[Claim 6]

The method of claim 5,

(f) generating a node item including the location and size of the node entry created in step (e) and storing the node item in the node page.

[Claim 7]

The method of claim 6,

In the step (f),

And each node item corresponding to each node entry is stored on the node page by sorting the data based on the index key column included in the corresponding node entry.

[Claim 8]

(a) receiving a mixed query for data retrieval;

(b) selecting an index most closely related to all of the columns included in the blend query;

(c) a first node entry and a last node entry including index key data identical to the index key data included in the mixed query among the node entries included in the selected index using the index key meta information of the selected index. Confirming;

(d) extracting a node entry containing the same non-index data as the non-index data included in the mixed query while moving from the identified first node entry to the last node entry; Treatment method.

[Claim 9]

(a) receiving a mixed query for data retrieval;

(c) Node entry identification included in the selected index using index key meta information of the selected index. First node entry and last node entry containing the same index key data as the index key data included in the mixed query. Confirming;

(d) extracting non-index data corresponding to the non-index column included in the mixed query from each node entry moving from the identified first node entry to the last node entry. Treatment method.

[Claim 10]

A computer-readable recording medium having recorded an index data structure including at least one index node for performing data retrieval,

Each of the index nodes includes at least one node entry which is a basis of data retrieval, and each node entry includes a plurality of storage areas each storing data extracted from different columns of any one record in a database table. Computing the index data structure characterized in that Recordable media.

[Claim 11]

The method of claim 10,

At least one of a plurality of storage areas included in each node entry stores index key data in an index key column of the table, and the non-index column of the table is stored in the remaining storage areas of the plurality of storage areas. A non-indexable data stored on a computer readable recording medium having recorded thereon an index data structure.

[Claim 12]

The method of claim 10,

Each storage area for storing data extracted from each column includes a data storage area for storing real data and a data information storage area for storing the location and size of the data storage area. Readable media.

[Claim 13]

The method of claim 10

And wherein each index node includes at least one fixed size node item in which the location and size of each node entry is stored.

[Claim 14]

14. The method of claim 13, ^'

A computer-readable recording medium recording an index data structure, wherein the total size of each node entry depends on the number of columns stored and the size of the actual data.