This article includes suggestions and workarounds. Content may not be accurate for all use cases or represent best practices for the latest release.
Why are my tasks sometimes failing due to missing mappings and not importing data, while the data is imported when running directly through the loader?
Tasks can include multiple loaders. When a loader fails within a task, subsequent loaders in the same task will not process as we do not want to load any incomplete or conflicting data. The earlier loader may have failed due to missing mappings. You can find more specifics on missing mappings below.
When you run a loader manually, and there are missing mappings for levels or accounts, the data that is mapped will be imported, and an error will be shown in the logs stating the mappings that are missing. This is true for all standard account imports, as well as cube sheet imports. Modeled sheet imports will be successful without account mappings, and fail without level mappings.
When a task is run, it is either marked as a success, or a failure. A task may fail due to missing account or level mappings, but the mapped data will still be imported for the loaders that were run within the task. If a task has multiple loaders, you can check the task logs to identify which loaders have been run, and confirm the error. Similar to the loader, the task will be a success or failure depending on the type of import it is. Cube and Standard act the same, while Modeled imports do not require accounts.
If a task or loader has missing dimension mappings, they will be classed as a success because dimension values are not necessary, and data can still be imported to the account and level combination. This may be why a task is successful when showing missing mappings.