5.2.2.7 Managing Data Import

You can control the priority of ETL as well as the amount of I/O resources that ETL consumes using I/O Resource Management (IORM).

Data import, or extract, transform, load (ETL), is an important part of maintaining a data warehouse. In some cases, ETL is extremely critical to performance because reports or queries cannot be run until the data has been loaded. In these cases, ETL should be prioritized above all other queries. In other cases, ETL is a low-priority background activity that only needs to be prioritized in the rare event that it does not complete by a certain time.

To manage ETL, do the following:

• Map the ETL sessions to the ETL_GROUP consumer group.

  The mapping rules for ETL are typically based on user name or client program name. Data pump is run under the DATALOAD function. By default, the DATALOAD function is mapped to the ETL_GROUP consumer group.

• Include the ETL_GROUP group in your resource plans.

To import non-compressed data as compressed data,

Example 5-8 Mapping a Program to the ETL_GROUP Consumer Group

This example shows how to map a program to the ETL_GROUP consumer group.

BEGIN
DBMS_RESOURCE_MANAGER.CREATE_PENDING_AREA();
DBMS_RESOURCE_MANAGER.SET_CONSUMER_GROUP_MAPPING  
  (DBMS_RESOURCE_MANAGER.CLIENT_PROGRAM, 'SQLLDR', 'ETL_GROUP');
DBMS_RESOURCE_MANAGER.SUBMIT_PENDING_AREA();
END;
/

Importing Non-compressed Data as Compressed Data

Non-compressed data can be imported as compressed data when using the TRANSFORM:SEGMENT_ATTRIBUTES=n option, and the target tablespace has been configured to create new tables as Exadata Hybrid Columnar Compression tables by default.