The Bad Record Exception examines input record scores to determine record quality. It returns records to different output groups
The Exception transformation identifies the following types of records based on each record score:
Good records
Records with scores greater than or equal to the upper threshold. Good records are valid and do not require review. For example, if configure the upper threshold as 90, any record with a 90 or higher does not need review.
Bad records
Records with scores less than the upper threshold and scores greater than or equal to the lower threshold. Bad records are the exceptions that you need to review in the Analyst tool. For example, when the lower threshold is 40, any record with a score from 40 to 90 needs manual review.
Rejected records
Records with scores less than the lower threshold. Rejected records are not valid. By default, the Exception transformation drops rejected records from the data flow. For this example, any record with a score 40 or less is a rejected record.
If the quality issues fields are NULL, the record is not an exception. When any quality issue contains text or it contains an empty string, the record is an exception. Verify that a quality issue port contains null values when a field has no error. If the quality issue ports contain blanks instead of null values then the Exception transformation flags every record as an exception. When a user needs to correct the issues in the Analyst tool, the user cannot filter the exceptions by data quality issue.
When a record has a score less than zero or greater than 100, the row is not valid. The Data Integration Service logs an error message that the row is not valid and it skips processing the record.
If you do not connect a record score as input to the Exception transformation, the transformation writes all records that contain quality issues to the bad records table.
When you include the Bad Record Exception transformation in a Mapping task, you can configure a Human task in the same workflow to include a manual review of the exceptions. The Human task starts when a Mapping task ends in the workflow. The Human task requires users to access the Analyst tool to resolve the quality issues. A user can update the data and change the quality status of each record in the bad records table.