Multi-Source Execution Plans

Best Practices for Multi-Source Execution Plans

The following rules apply to homogeneous and heterogeneous multi-source execution plans.

Phase Dependency Blocking Behavior

DAC’s Task Phase Dependency feature enables you to change the order in which tasks are executed.

Task phase dependencies have a Scope property, which specifies how the Block action behaves in relation to multi-source execution plans.

For more information about phase dependencies, see "Tasks Tab: Phase Dependency Subtab".

You can set the Scope property to one of the following values:

Both

Indicates the blocking action is active for tasks that have the same source and

target physical data source connections.

Source

Indicates the blocking action is active for tasks that have the same source physical data source connection.

Target

Indicates the blocking action is active for tasks that have the same target physical data source connection.

None

Indicates the blocking action is active for all tasks regardless of the source and

target physical data source connections.

Delay Property

The Delay property is located in the Parameters sub tab in the Execution Plans tab of  the Execute view.

You set this property to specify how many minutes an extract of a data source will be delayed after the first extract of a multiple source extract process started.

Source Priority

For multi-source execution plans, you must assign each physical data source a priority (in the Physical Data Sources tab of the Setup view). The priority ranks the data sources and specif ies the order in which DAC will load the data from the different  sources. This property ensures that tasks attempting to write to the same target table will not be in conflict.

Truncate Table Behavior

If a target table is shared across different sources, it will be truncated only once. The priority of the data source determines which of the extracts truncates the tables. The task reading from the data source with the highest priority truncates the tables and drops the indexes. The last task writing to the table from the data source with the highest priority creates the indexes.

Truncate Options

Truncate options should be the same across the source system containers.

Task Groups

Tasks should belong to the same task groups across the source system containers.

Customized Tasks for Vertical Applications

In the case of Vertical Applications in which a task could have been customized in a

Vertical workflow folder, the task will pick up the Vertical task folder when the source

and target connections are the same based on the task folder priority.

DATASOURCE_NUM_ID

■ All tables should contain the column DATASOURCE_NUM_ID.

■ Unique indexes should always include the column DATASOURCE_NUM_ID.

■ All the extract mappings populate the DATASOURCE_NUM_ID column from the

parameter file produced by DAC.

■ All the load mappings extract the value of the DATASOURCE_NUM_ID column

from the staging tables.

■ Level 1 aggregate tables should always contain the column DATASOURCE_

NUM_ID.

Load Strategies

You should use the same load strategy for loading a table across the source system

containers; that is, you should always use full loads or always use incremental loads.

Avoid using a combination of load types.

Failure Restart Strategies

The Upon Failure Restart option of the Task Actions feature can be useful to restart

tasks upon failure. For information about using task actions, see "About Index, Table

and Task Actions".