KR100238439B1 - Method of managing object-orient route index of schema manager - Google Patents

Abstract

The present invention relates to a method of managing a path index in an object-oriented kernel of an object-oriented DBMS, although there is conventionally supported a path index in an object-oriented DBMS, but there is no management method registered, the present invention to solve the above problem. Supports multiple indexes to speed up database searches for queries using path expressions in object-oriented DBMSs. When searching and managing information needed for multiple indexes using a catalog, the search speed during querying and duplicate index creation are improved. By proposing a preventive catalog structure and management method, it is possible to prevent the creation of duplicate indexes for the same attribute when creating multiple path indexes.

Description

Object-Oriented Path Index Management in Schema Manager

The present invention relates to a method for managing a path index in an object-oriented kernel of an object-oriented DBMS, although there is a support for the path index in an existing object-oriented DBMS, but a management method thereof is not registered as a patent.

The present invention supports multiple indexes to speed up the database search for a query using a path expression in an object-oriented DBMS. When the information necessary for managing multiple indexes is managed using a catalog, the search speed is improved and a duplicate index is used. I would like to suggest a catalog structure and management method to prevent creation.

1 is a hardware configuration diagram to which the present invention is applied;

2 is a module configuration diagram of the object-oriented kernel of the present invention;

3 is a structural diagram of a catalog class managed by a schema manager of an object-oriented kernel of the present invention;

4 is an operation flowchart when creating an index in the object-oriented kernel according to the present invention;

5 is an operation flowchart of a schema manager when a path index is generated in an object-oriented kernel according to the present invention;

6 is a flowchart illustrating an index checking operation for each attribute when processing a query using a path expression in the object-oriented kernel according to the present invention.

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

1 is a hardware configuration diagram to which the present invention is applied.

The configuration includes a processor 11 with Pi (1 <= i = n), a main memory area (Mem) 12, an input / output dedicated processor (IOP) 13, and a disk (DISK) 14 as an auxiliary memory. )

The hardware can operate in a general hardware environment including a single processor or multiple processors and memory regions.

2 is a module configuration diagram of the object-oriented kernel of the present invention.

2 shows a module structure of an object-oriented kernel in an object-oriented DBMS in which the present invention operates.

In view of the above structure, the object-oriented kernel to which the present invention is applied basically includes an interface management module 22 connecting the user 21 and the system, a schema management module 23 managing the catalog and the user-defined class, and a query. It consists of a query processing module 24 for executing the and the object management module 25 for managing all objects.

Data storage in the database 27 uses the data storage management system 26.

3 is a structural diagram of a catalog class managed by the schema management module of the object-oriented kernel of the present invention.

The catalog class includes SM_CLASS (31) for managing class information, SM_ATTRIBUTE (32) for managing information on each attribute defined in the class, and SM_METHOD (33) for managing information for each method defined in the class. SM_INDEXS (34), which manages each index information of the class, SM_KEY_DESC (35), SM_DICTIONARY (36), which manages dictionary information used for the specialized information index, and SM_PRIVILEGE (, which manages the user's authority information for the class. 37) and SM_USER (38) which stores registered user information, and provides all schema related functions by operating on them.

The catalog class structure related to the present invention is SM_CLASS (31), SM_ATTRIBUTE (32), SM_INDEX (34), and SM_KEY_DESC (35). In particular, SM_ATTRIBUTE and SM_KEY_DESC are used as follows.

The index information required for query processing in the object-oriented kernel should be known from the attributes used in the query, and maintains the index list in SM_ATTRIBUTE whose index is the first key in the index_list field.

If you do not maintain this, all index information defined in the class should be obtained from index_list of SM_CLASS, and it will be slow.

In addition, in the case of a path index when an index is created, if there are the same attributes included in several paths at the same time, an index using the attribute may be created many times. In order to prevent waste of other time and storage place in such duplicate creation, SM_KEY_DESC (35 ) Maintains the number of index references in the index_count field.

In other words, when there is a need to create an index such as an index that has already been created, the number of index references is increased to prevent duplicate creation.

4 is a flowchart illustrating an operation of creating an index in the object-oriented kernel according to the present invention.

First, when the request (41) to use the index from the user, the schema management module catalogs the information necessary for creating the index (class name to create the index, attribute name used as the index key) in the SM_CLASS (31) and SM_ATTRIBUTE (32) catalogs. Check using the information of (42).

After the check, if appropriate, a new index identifier is generated (43), and an index is created using a function of the object management module in a class extent in which objects of the corresponding class exist (44).

If the index creation is successful, index information necessary for the SM_CLASS, SM_ATTERIBUTE, SM_LNDEX, and SM_KEY_DESC catalog classes is stored (45).

5 is a flowchart illustrating an operation of a schema manager when a path index is generated in an object-oriented kernel according to the present invention.

Initially, information about a class to be indexed is obtained from SM_CLASS (51).

From the class information, an attribute list (attribute_list), a next index number (next_index_id) of a class, an index list (index_list) and a class identifier (class_id) already existing in the class are obtained.

The instance of the class is obtained using the obtained class identifier.

Information about all key attributes in the path of the path index is obtained from the SM_ATTRIBUTE (52).

In the above, the first attribute is one of the attributes of the current class, and the second is the attributes of another class.

From the attribute information, the attribute type (attr_type) and index information (index_list) having the attribute as a key are obtained.

Next, it is determined whether it is a final key attribute (53).

In the case of the last attribute after the above determination, the attribute type has a base type instead of another class, or it should always have a different class type.

If it is the last attribute after the last attribute determination (53) (512), the index information in the attribute information is obtained from SM_KEY_DESC (56).

If it is not the last attribute after the determination (53) (513), it is determined whether the attribute type is a class type (54), if it is not a class type (515), and error processing (55) ends because it is an error (514). ) Next index information is obtained from SM_KEY_DESC (56).

The index information includes key attributes of the index (key_attributes), the number of index references (index_count), the index number (index_id), and the like.

In operation 57, it is determined whether an index using the key attribute of the index information already exists.

The presence or absence of an index indicates that the key attribute (key_attributes) of the index information indicates the information about the attribute (SM_ATTRKBUTE), and the index information (index_list) can be immediately known from the attribute information.

If the same index exists (516), the number of references (index_count of SM_KEY_DESC) of the index information is increased by one without generating additional indexes (511).

This prevents duplicate occurrences of the same index, reducing execution time and storage location.

If the same index does not exist (517), the next index number to be created is assigned (58).

The next index number is kept in the class information (nex_index_id of SM_CLASS) and after getting it, increment one and store it again.

After storing, an index is created in the corresponding class extent (59).

You can easily find class extents by keeping the identifier of the class in the attribute and index information.

Various information of the generated index is stored in the SM_CLASS, SM_INDEX, SM_ATTRIBUTE, SM_KEY_DESC catalog (510).

The stored index information is used to create another index or to create an optimal execution plan when processing a query, or retrieve the index information when a user requests the index information.

After storing various information of the index in the SM_CLASS, SM_INDEX, SM_ATTRIBUTE, SM_KEY_DESC catalog, the number of index references of the index information is increased (511).

After generating the index or increasing the reference number of the index information, it is again performed from the determination of the last key attribute to generate an index for the next attribute of the path (518) and ends if it is the last attribute.

6 is a flowchart illustrating an index checking operation for each attribute when processing a query using a path expression in the object-oriented kernel according to the present invention.

First, an index list (index_list of SM_ATTRIBUTE) having the attribute as a key is obtained from the attribute information (62).

The index list includes object identifiers for index information as elements.

By pointing the index information directly in the attribute information as in the above procedure, it is faster than finding and comparing all the indexes defined in the class.

After obtaining the index list, index information SM_KEY_DESC is obtained using the index object identifier (3).

It is determined whether the current attribute information is the same object based on the attribute object identifier included in the key attribute list (key_attributes of SM_KEY_DESC) in the index information (64).

By maintaining the object identifier for the key attribute in the index information, it is checked whether the attribute used in the query is the same through the object identifier comparison, so the speed is faster as the attribute information does not need to be obtained and compared.

If it is the same (610), the index information means the corresponding index that can be used in the current path expression, so that the query processing module is informed of the index information using the current attribute as a key (67).

If it is not the same after the check (611), it is determined whether the last element of the index list (65).

If it is not the last element after the determination (612), the index information is obtained by using the index object identifier again to repeat the index check until the last element.

If it is the last element of the list (613), it informs the query processing module that no index has been created for the current attribute as a key (66).

In the present invention, in order to support multiple indexes for quickly searching a database for a query using a path expression in an object-oriented DBMS, the catalog presented when managing information necessary for managing multiple indexes in a schema manager of an object-oriented kernel using a catalog The structure and management method improves the search speed when querying and prevents duplicate index generation for the same attribute when creating multiple path indexes.

An object of the present invention is to support multiple indexes, which are object-oriented path indexes, as a catalog structure and a management method that can improve search speed and prevent duplicate index generation in a schema manager of an object-oriented kernel.

Claims (3)

The interface management module 22 which connects the user 21 and the system, the schema management module 23 which manages the catalog and user-defined classes, the query processing module 24 which executes the query, and all the objects. In object-oriented index management consisting of an object-oriented kernel composed of an object management module 25 and storing data in a database 27 using the data storage management system 26, SM_CLASS (31), which manages class information, contains information necessary for index creation (class name to create index, attribute name to use as index key) in the schema management module when request 41 for creating an index comes from the user. A first step of identifying (42) using information of an SM_ATTRIBUTE (32) catalog that manages information about each attribute defined in (42); Generating a new index identifier (43), if appropriate, and creating an index (44) using a function of an object management module on a class extent in which objects of the class exist; When the index generation is successfully performed, SM_CLASS (31) managing the class information, SM_ATTRIBUTE (32) managing information on defined attributes, SM_INDEX (33) managing each index information of a class, and SM_KEY_DESC (35). And a third step of storing the index information necessary for the catalog class (45). The method of claim 1, wherein the operation of the schema manager when generating the path index comprises: (51) obtaining information about a class to generate an index from the SM_CLASS; Obtaining, from the class information, an attribute list of a class, a next index number, an index list and a class identifier already present in the class; Obtaining an extent of the corresponding class using the obtained class identifier; Obtaining (52) information about all key attributes in the path of the path index from SM_ATTRIBUTE; A fifth step (53) of determining whether it is the last key attribute after obtaining information on the key attribute; Determining whether the type of the attribute is a class type if it is not the last key attribute (54) and performing an error processing (55) because the error is not a class type; A seventh step of obtaining (56) the next index information from the SM_KEY_DESC if it is the last key attribute in the fifth step or is not an error in the sixth step; An eighth step of determining (57) whether an index using the attribute is a key by checking a key attribute of the index information; A ninth step of assigning a next index number to be generated if the same index does not exist in the eighth step (58), and maintaining the next index number in the class information, and incrementing and storing the next index number again; A tenth step of maintaining an identifier of the class in the attribute and the index information by generating an index in the corresponding class extent after the storing; An eleventh step of maintaining an identifier of the class in the attribute and the index information by generating an index in the corresponding class extent after the storing; A twelfth step of storing (510) various information of the generated index in a catalog; In the eighth step, the same index information is present, or after performing the twelfth step, the number of index references of the index information is increased (511). A thirteenth step; And generating a index for the next attribute of the path after the index is created, and performing the determination of the last key attribute and ending if the last attribute is completed. The method of claim 1, wherein the index identification for each attribute in the query processing comprises: (62) a first step of obtaining an index list having the attribute as a key from attribute information; (63) a second step of obtaining index information using the index object identifier after obtaining the index list, and determining whether the current attribute information is the same object using the attribute object identifier included in the key attribute list in the index information (64); With three steps; (69) a fourth step of determining if it is not equal after the checking (69) is the last element of the index list; If it is not the last element after the determination, the index checking is repeated until the last element (610), and if the last element (611), the query processing module notifies that the index with the current attribute is not generated (66). Steps; (68) a sixth step of passing the index information, which is the key of the current attribute, to the query processing module if it is the same after the determination of the third step; and (68) a method of managing the object-oriented path index of the schema manager. .