This is a good strategy if there are clearly defined expected errors in external data, and if they are likely to affect only a small subset of the data.
To detect and repair expected import errors before bulk import:
1. | Define and create a reject table if necessary. |
2. | Define an error import task importing records that have expected errors into the reject table (use a WHERE clause to test for the error). |
3. | Define the main import task. This job can be based on the assumption that the expected error(s) no longer exist(s) in the data. |
| Alternatively, you can append the main import task to the error import task, and make it conditional upon the error import no longer affecting any data. |
4. | Run the error import task. |
5. | Inspect the reject table. On the basis of this information, fix the errors in the external file. |
6. | If necessary, run the error import task again to test that the error no longer occurs. |
7. | Run the main import task. |
|