90 DBMS_HS_PARALLEL

The DBMS_HS_PARALLEL PL/SQL package enables parallel processing for heterogeneous targets access. This package is designed to improve performance when retrieving data from a large foreign table.

This chapter discusses the following topics:

90.1 Using the DBMS_HS_PARALLEL Subprogram

DBMS_HS_PARALLEL is compiled with the authorization ID of CURRENT_USER, which uses invoker's rights. In other words, all procedures in this package are executed with the privileges of the calling user.

90.2 Summary of DBMS_HS_PARALLEL Subprograms

This table lists the DBMS_HS_PARALLEL subprograms in alphabetic order and briefly describes them.

Table 90-1 DBMS_HS_PARALLEL Package Subprograms

Subprogram Description

CREATE_OR_REPLACE_VIEW

Creates (or replaces) a read-only view to be referenced for retrieving the data from a remote table in parallel.

CREATE_TABLE_TEMPLATE

Writes out a CREATE TABLE template based on information gathered from the remote table. You can use the information to add any optimal Oracle CREATE TABLE clauses.

DROP_VIEW

Drops the view and internal objects created by the CREATE_OR_REPLACE_VIEW procedure. If the view has not already been created by the CREATE_OR_REPLACE_VIEW procedure, an error message is returned.

LOAD_TABLE

Loads the data from a remote table to a local Oracle table in parallel. If the local Oracle table does not already exist, it is created automatically.

90.2.1 CREATE_OR_REPLACE_VIEW Procedure

This procedure creates (or replaces) a read-only view to be referenced for retrieving the data from a remote table in parallel.

Syntax

CREATE_OR_REPLACE_VIEW (remote_table, database_link, oracle_view, parallel_degree)

Parameters

Table 90-2 CREATE_OR_REPLACE_VIEW Procedure Parameters

Parameters Description

remote_table

The name of the remote database table. It is specified as [remote_schema_name.]remote_table_name.

database_link

The remote database link name. The call can only be applied to a heterogeneous services database link.

oracle_view

The name of the Oracle view. It is specified as [schema_name.]oracle_view_name. The default schema name is the current user. If the oracle_view parameter is not specified, the remote table name will be used as the view name.

parallel_degree

The number of parallel processes for the operation is computed based on the range-partition number if applicable, or the number of CPUs. The range of values is 2 to 16.

Usage Notes

  • The specified Oracle view is created and future reference of this view utilizes internal database objects for parallel retrieval of remote non-Oracle table data. If the Oracle view already exists, the following Oracle error message is raised:

    ORA-00955: name is already used by an existing object
    
  • This view is created as a read-only view. If you attempt to insert and update the view, the following Oracle error message is raised:

    ORA-01733: virtual column not allowed here
    
  • If the remote table or the database link does not exist, one of the following Oracle error messages is raised:

    ORA-00942: table or view does not exist
    or
    ORA-02019: connection description for remote database not found
    
  • You need the CREATE VIEW, CREATE TABLE, CREATE TYPE, CREATE PACKAGE, and CREATE FUNCTION privileges to execute the CREATE_OR_REPLACE_VIEW procedure.

  • If you encounter either of the following Oracle error messages, increase the PROCESSES and SESSIONS parameter in the Oracle initialization parameter file:

    ORA-12801: error signaled in parallel query server P003
    or
    ORA-00018: maximum number of session exceeded
    
  • Because the CREATE_OR_REPLACE_VIEW procedure creates some internal objects, use the DROP_VIEW procedure to drop the view and the internal objects. The SQL DROP VIEW statement only drops the view and not the internal objects.

90.2.2 CREATE_TABLE_TEMPLATE Procedure

This procedure writes out a CREATE TABLE template based on information gathered from the remote table. You can use the information to add any optimal Oracle CREATE TABLE clauses.

Syntax

CREATE_TABLE_TEMPLATE (remote_table, database_link, oracle_table, create_table_template_string)

Parameters

Table 90-3 CREATE_TABLE_TEMPLATE Procedure Parameters

Parameter Description

remote_table

The name of the remote database table. It is specified as [remote_schema_name.]remote_table_name.

database_link

The remote database link name. The call can only be applied to a heterogeneous services database link.

oracle_table

The name of the local Oracle table the data will be loaded into. It is specified as [schema_name.]oracle_table_name. The default schema name is the current user. If the oracle_table parameter is not specified, the remote table name will be used as the local Oracle name.

create_table_template_string

Contains the Oracle CREATE TABLE SQL template when the procedure is returned.

90.2.3 DROP_VIEW Procedure

This procedure drops the view and internal objects created by the CREATE_OR_REPLACE_VIEW procedure. If the view has not already been created by the CREATE_OR_REPLACE_VIEW procedure, an error message is returned.

Syntax

DROP_VIEW (oracle_view)

Parameters

Table 90-4 DROP_VIEW Procedure Parameter

Parameter Description

oracle_view

The name of the Oracle view created by the CREATE_OR_REPLACE_VIEW procedure. If the view has not been created by the CREATE_OR_REPLACE_VIEW procedure, an error is returned.

90.2.4 LOAD_TABLE Procedure

This procedure loads the data from a remote table to a local Oracle table in parallel. If the local Oracle table does not already exist, it is created automatically.

Syntax

LOAD_TABLE (remote_table, database_link, oracle_table, truncate, parallel_degree, row_count)

Parameters

Table 90-5 LOAD_TABLE Procedure Parameters

Parameter Description

remote_table

The name of the remote database table. It is specified as [remote_schema_name.]remote_table_name

database_link

The remote database link name. The call can only be applied to a heterogeneous services database link.

oracle_table

The name of the local Oracle table the data will be loaded into. It is specified as [schema_name.]oracle_table_name. The default schema name is the current user. If the oracle_table parameter is not specified, the remote table name will be used as the local Oracle name.

truncate

Determines whether the Oracle table is truncated before the data is loaded. The value is either TRUE or FALSE. The default value is TRUE which means the Oracle table is truncated first. When set to FALSE, the Oracle table will not be truncated before the data is loaded.

parallel_degree

The number of parallel processes for the operation is computed based on the range-partition number if applicable, or the number of CPUs. The range of values is 2 to 16.

row_count

Contains the number of rows just added with the load table operation.

Usage Notes

  • This procedure only loads the remote table data into Oracle local table. It does not create a key, index, constraints or any other dependencies such as triggers. It is recommended that you create these dependencies after the table data is loaded as performance will improve greatly. You will need to decide whether to create the dependencies before or after the data is loaded based on your knowledge of the remote table data and dependencies.

  • If the local table does not exist, the LOAD_TABLE procedure creates a simple (non-partitioned) local table based on the exact column matching of the remote table after which the data is inserted into the local table.

  • If the remote table or the database link does not exist, an error message is returned.

  • If the local table is incompatible with the remote table, an error message is returned.

  • You need the CREATE TABLE, CREATE TYPE, CREATE PACKAGE, and CREATE FUNCTION privileges to execute the LOAD_TABLE procedure.

  • If you encounter either of the following Oracle error messages, increase the PROCESSES and SESSIONS parameter in Oracle initialization parameter file:

    ORA-12801: error signaled in parallel query server P003
    or
    ORA-00018: maximum number of session exceeded
    
  • One of the following is required for parallel processing:

    • The remote table is range partitioned.

    • Histogram information for a numeric column is available.

    • There is a numeric index or primary key.

  • To drop the local table, use the DROP TABLE SQL statement.