We can sometimes create editable tables in "SQL flows" (e.g. to create mappings between codes and their corresponding meaning).
However, the editable table can not be used by SQL recipes because its type does not conform to the expected format (see below screenshot showing the shaded SQL recipe, at our organization SQL recipes use AWS Athena):
Note: SparkSQL recipes do work on editable tables but at the expense of facing Spark's overhead (so one can wait several minutes to synchronize a very small table).
In this case, a workaround is to convert the editable table in parquet, for example by using a sync recipe, so that SQL recipes can now read it (see below screenshot where the SQL recipe is now available):
It would make things simpler if dataiku directly offered the possibility to save the editable table in the desired format (in my case on S3 if I am correct).