Thursday, October 02, 2008
Exception Handling is necessary evil, which is required for every application. But if it is not done properly with appropriate design and requirement in mind, it can be one of the biggest challenge in maintaining the application in long term. The approach for exception handling should be
1. No change in data by default
2. provide easy access and traceability of error to source system
3. process in place to notify business user
4. if no action taken for specific period of time, archive and send notification
Exception Handling: Exceptions are defined in two categories namely Critical and Warning. Critical exception are those which stops the incoming row from going to business target table. Instead row information is captured in exception table(s) and presented to business team to analyze the error. Warnings exceptions are those business rule failure which does not stop the row from moving to business target table(s) but some exception warning information is captured in exception table(s) for later review.
As in any other tool, error can be pragmatically captured in Informatica. But Informatica also provide inbuilt capability to capture error. This need some architectural design and development standard for exception handling. Informatica provide four tables ( or a file) where all error can be captured. Post process and use these tables (or file) to parse the error information and load into exception data model. This will provide standard method of error capture.
1. No change in data by default
2. provide easy access and traceability of error to source system
3. process in place to notify business user
4. if no action taken for specific period of time, archive and send notification
Exception Handling: Exceptions are defined in two categories namely Critical and Warning. Critical exception are those which stops the incoming row from going to business target table. Instead row information is captured in exception table(s) and presented to business team to analyze the error. Warnings exceptions are those business rule failure which does not stop the row from moving to business target table(s) but some exception warning information is captured in exception table(s) for later review.
As in any other tool, error can be pragmatically captured in Informatica. But Informatica also provide inbuilt capability to capture error. This need some architectural design and development standard for exception handling. Informatica provide four tables ( or a file) where all error can be captured. Post process and use these tables (or file) to parse the error information and load into exception data model. This will provide standard method of error capture.
Labels: Exception Handling using Informatica
# posted by Arvind @ 10:50 AM 
