Schema changes in the data sources

If some of the recipes of the "data extracts " Flow zone fail, it is certainly due to a change of schema in one of the sources. It can especially happen on the Big Query Data Lake sources as they might evolve over time.

To resolve it, access the failing dataset and go in the "Settings" tab. From here, you can access the "Schema" tab, check it and update the schema from the source table if needed.

Once done, access the following Sync recipe and rclick the "Resync schema" button, then click "Run" at the bottom left of the screen.

Imputer error

The following error can happen if the imputer is called on an empty dataset. The imputer being called product family by product family, this can happen if one of the family does not have any CPC.

This probably means there is another issue at a previous stage of the Flow that went undetected. In this case, find the root of the issue in the data preparation steps.

StackOverflowError

This error happens when Dataiku fails at computing the dependencies between all of the elements.

To avoid this, when building a dataset it is advised to select the "Not recursive" option instead of the "Recursive upstream" one.

Best params dataset is not ready

When creating a new branch for a development on one of the projects, it is necessary to rebuild some tables as shown here (example for SpP GBU). This error will appear if they are not rebuilt before running the scenarios.

Checks failing

If any check is failing, please refer to the dedicated page to better understand how to resolve the situation.