While defining mappings from a source schema to a target schema, a few mappings in a transformation might have resulted in warnings or errors. A transformation can be deployed to the Administrative Service if it has warnings, but it cannot be deployed if it has errors. In case of a warning, the result of the transformation at the Data Exchange Runtime depends on the actual data being transformed. For example, when you define a mapping from a real (RDMS) to an int (SQL Server), a warning is generated. If the actual source data is 500, then the data is transformed into the int type without any issue. However, if the actual source data is 500.35, then a Runtime error will be encountered when this data is being transformed. When you define the transformation in DDW, you can choose how you want the Data Exchange Runtime to handle this kind of data.
You must configure the data transform error handling policies before deploying transformations to the Administrative Service. You can define a policy at a transformation level or at a classifier level.
Notes:
Although the policy is set in DDW, it is effective in Runtime during data transformation.
The policies are set to handle data conversion, filter, and DML command substitution errors.
By default, the policy at the transformation level is set to Stop.
All policies defined at the transformation level are automatically applied to each classifier level.
The policy defined at the classifier level overrides the policy defined at the transformation level.
The policies are applicable to both CDT and BDT.
The following links provide more information about configuring the error handling policy at the transformation and classifier levels.