Transaction Processing and JSON Database Workloads with Autonomous AI Database

Autonomous AI Database configures and optimizes your database for you, based on your workload.

Characteristics of Autonomous AI Database with Transaction Processing or JSON Database workloads:

• The default data and temporary tablespaces for the database are configured automatically. Adding, removing, or modifying tablespaces is not allowed. Autonomous AI Database creates one tablespace or multiple tablespaces automatically depending on the storage size.

• The database character set is Unicode AL32UTF8. See Choose a Character Set for Autonomous AI Database for more information.

• Compression is not enabled by default but Autonomous AI Database honors a compression clause if compression is specified on a table.

Accessing a database:

• You do not have direct access to the database node. You can create and drop directories with CREATE DIRECTORY and DROP DIRECTORY, as described in Create and Manage Directories.

  You can use DBMS_CLOUD procedures such as DBMS_CLOUD.DELETE_FILE, DBMS_CLOUD.GET_OBJECT, and DBMS_CLOUD.PUT_OBJECT with files and objects. You do not have direct access to the local file system.

Parallel Execution with Transaction Processing or JSON Database workloads:

• Parallelism is determined by the database service you use. See Database Service Names for Autonomous AI Database for details about parallelism support for each database service.

• When you want to run DML operations in parallel and the database service you are using allows this, you can enable parallel DML in your session using the following SQL command:

  ALTER SESSION ENABLE PARALLEL DML;

  See VLDB and Partitioning Guide for more information on parallel DML operations.

• If you create an index manually and specify the PARALLEL clause, the PARALLEL attribute remains after the index is created. In this case SQL statements can run in parallel unbeknownst to the end user.

  To specify serial execution, change the INDEX parallel clause to NOPARALLEL or set the PARALLEL degree attribute to 1 to specify serial execution:

  ALTER INDEX index_name NOPARALLEL;

  or

  ALTER INDEX index_name PARALLEL 1;