Task Groups

Best Practices for Task Groups

The following best practices apply to task groups:

1)    Do not create task groups unnecessarily, because this can impact a better way of ordering tasks.

2)    Create task groups only if you want the following scenarios and the auto-dependency (auto ordering of tasks for an execution plan) cannot solve the order in a proper way.

3)     Avoid circular relationships among tasks

. For example, avoid situations in which Task 1 reads from Table A and writes to Table B, and Task 2 reads from Table B and writes to Table A. You can use task groups to avoid these situations. If the tasks belong to different phases, however, this situation is acceptable.

4)     If you have many tasks belonging to the same phase that write to the same table, you can use task groups to run the tasks in parallel.

5)    If the target tables need to be truncated before the tasks are run, select the properties Truncate Always and Truncate Full Load in the Task Group tab.

■ Do not mix extracts and loads under a single table group.

■ Do not make Task Groups for obvious ordering needs. DAC handles ordering in such cases.

■ If a source system container uses a task group, make other containers that

reference the task also include the task group.