A session is a set of instructions that tells the Informatica server how and when to move data from sources to targets. You create a session for each mapping that you want the server to run. If you wish, you can group multiple sessions together and run them in batches. Batches provide a way to group sessions for either serial or parallel execution by the Informatica server; see the Glossary for more information. All the sessions in a batch must occur in the same folder. If you want to group multiple sessions together but they occur in different folders, you need to use the PMCMD utility rather than a batch.

You use Informatica’s Server Manager to create and maintain sessions. You can schedule a source-to-target mapping to run on a scheduled or ad hoc basis.
When you schedule a session, you must specify:

• The type of load (Insert/Update/Delete/Data Driven)

• Source and target connection information. The source can be a flat file or a database. You can truncate the target files in case of errors, if you wish.

• Any desired pre- and post-session routines

• Memory allocation and buffer size

• Load schedule

• Log file location

• Transformation overrides

Sessions require:

• One or more source tables or files

• One or more target tables

• One or more transformation objects

• A Source Qualifier transformation is always required in order to extract the data from the source (except for COBOL sources, which always use the Normalizer transformation to extract data).

Informatica’s Session Wizard takes you through the process of creating and scheduling a session. You start the Session Wizard in the Server Manager. The Session Wizard has five main windows:

• Mappings

• General

• Time

• Log Files

• Transformations

In the Session Wizard’s Mappings window you select the name of the mapping for which you want to create a session.

The Session Wizard’s General window automatically names the session to s_mappingname, where mappingname is the name of the mapping for which you are creating a session. In the General Window you specify source and target information, session commands, and performance options, and choose whether to enable or disable the session.

In the Session Wizard’s Time window you schedule the session.

In the Session Wizard’s Log Files window you specify options related to log files and error handling. Each session produces a log file. The log files provide detailed information at a system level. Information on errors in the data load process itself is written to error tables in the Connect database.

Log file options:

• Saving the session log files from the last x runs will keep a rolling history of the last x log files (x is a number you specify). Older log files are deleted.

• Saving the session logs by timestamp adds a timestamp to each log file name. No logs are deleted.

Error handling options:

• Stop on x errors - Select this option if you want the loading process to fail when a certain number of errors occur (x is a number you specify).

• Perform recovery - This runs the session in recovery mode to start loading after a session failure.

• Override tracing - This logs the session in several levels of detail: Terse, Normal, Verbose Init, and Verbose Data.

• Batch Handling - Specify whether subsequent sessions should run if you are running sessions in a batch and one session fails.

• On pre-session command errors - Specify whether to stop the session if pre-session command errors occur.

• On stored procedure errors - Specify whether to stop the session if stored procedure errors occur.

In the Session Wizard’s Transformations window you can override transformation properties for the Source Qualifier, Lookup, Joiner, and Aggregator transformation objects.