Tasks

Best Practices for Tasks

The following best practices apply to tasks:

1)    Start your work with tasks in Informatica. After you create a workflow, do the following in the DAC Task tab:

– Create a new task and assign it a logical (readable) name.

– Enter the command for a full load or incremental load.

The commands can be the same.

If the Command for Incremental Load field is Left blank, no action occurs for this task while in incremental mode.

If the Command for Full Load field is left blank, no action occurs for this task while in full mode.

– Make sure all the source and target tables are defined for the task.

You can use the task synchronize functionality to import data from

Informatica. You can also manually assign the source or target tables.

2)    Select at least one primary table because the incremental and full mode properties are determined based on the refresh dates of the primary table.

3)  Design tasks so that they load only one table at a time.

4)  Define granular tasks rather than tasks that include bigger blocks of processes.

Granular tasks are more efficient and have better restartability.

5) Do not truncate a table on the source system tables (for example, Oracle, Siebel or Peoplesoft).

6)  Make sure the truncate property for the target tables is set properly.

7)  For tables that need to get truncated regardless of the mode of the run (Full or Incremental), set the Truncate Always property to True.

8)  For tables that need to get incrementally loaded, set the Truncate for Full Load property to True.

9)  Select the Analyze Table option if the task should analyze the table.

The default value for this option is True if either of the Truncate options are selected.

10)  Do not truncate a table more than once within the single life span of an ETL.

■ If a task that writes to a target table is contingent upon another table being loaded, use conditional tables. This ensures that the task qualifies only if the conditional table is part of the subject area design.

■ Assign an appropriate phase to the task. An understanding of task phases is

essential to understanding ETL processes.

■ If you want to force a relationship where none exists, consider using phase

dependencies.

For example, if you have an ETL process in which the extract facts and extract dimensions do not share any common source or target tables, but your design requires that the extract facts should run before extracting dimensions, then, for the task that extracts facts, add extract dimension as the phase that waits.

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

Dependency Subtab".

■ Make sure you do not introduce conflicting phase dependencies. This can cause

the DAC Server to hang.

■ If the source qualifier needs to use a data parameter, always use the DAC date

parameter that can be formatted to the database-specific syntax.

gn:n � t x ��� �� ce:none'>7)   The incremental load process requires additional logic to determine whether a record should be inserted or updated.

8)   Therefore, if you add a new source connection to populate an existing data warehouse, you should expect the incremental load to be slower than when running a full load.

 Incremental extract and incremental load

1)   This extract and load combination is used for regular nightly or weekly ETL processes.

2)   New or changed records are extracted from the source system and

          Loaded into the data warehouse.

3)   The DAC performs an incremental extract for a task if the source and staging tables have refresh dates and performs an incremental load for a task if the staging and target table have refresh dates.

i� + o h ��� �� e-font:major-latin;mso-bidi-font-family: Palatino-Roman'>Navigate to the appropriate folder, and select the authentication file. Click OK.

d. In the Configuring... dialog box, click Test Connection to confirm the

connection works.

e. Click Apply, and then click Finish.

Note: You must distribute this authentication file to all user accounts

that need to access this DAC Repository.

6. To create a new authentication file, do the following:

a. Click in the Authentication File field of the Configuring... dialog box.

b. In the Authentication File dialog box, select Create authentication file.

c. Navigate to the folder where you want to save the new authentication file, and

click OK.

d. In the Create Authentication File dialog box, enter a unique name for the

authentication file, and click OK.

e. Enter the Table Owner Name and Password for the database where the

repository will reside.

f. In the Configuring... dialog box, click Test Connection to confirm the

connection works.

g. Click Apply, and then click Finish.