Debugging Tips Errors in a transformation can be difficult to detect and resolve without proper debugging tactics. The first part of debugging is careful thought about each step in a transformation. Most steps operate like the exercise and data integration assignment. Focus on new steps for the case study as sources of errors. Sometimes, despite thinking carefully about each step, a transformation will terminate with errors (red lines in logging) or terminate with incorrect number of rows written to the output tables. An example of an error causing early termination is integrity errors in a table output step without error handling in the Table Output step. An example of incorrect number of rows is an incorrect filter that rejects more rows than a validation requirement. For both types of errors, you must determine the source of the error and correct the error. Follow these guidelines to determine the source (causation step) of the error. - To debug an error with early termination, determine a safe point in your transformation. A safe point is a step in which you are sure about correct processing. Disable all hops after the safe point and on alternative paths from the safe point. Preview the results of the safe point using the Preview button for the step. Move the safe point one step (enable hop to new step) to systematically find the step causing the error. Note that the error may not directly cause the failure in the step. The error may cause a failure in a later step. - For incorrect number of rows, look at the step metrics to determine the correct number of rows at each step. Preview the first step with incorrect number of output rows