Allow Data Fusions as ETL input datasets

Dean_Wangerin Minneapolis, MN 🟠

Currently Data Fusions are not allowed as inputs to ETLs. I'd like to propose that they be allowed as inputs, the same as any other dataset.


This is a gap because there's very little difference from a user point of view between fusions and other etls, but they behave differently.  A common pattern for power users is to build specialized ETLs off of general datasets.  But this isn't possible if the general dataset was created as a fusion.


Technically I've read that a datafusion is a view rather than a dataset itself. But, there's no reason that the ETL couldn't create a temporary table with the contents of the fusion and then run the ETL. This seems like a small technical hurdle for an easy win. 

4 votes

· Last Updated

This discussion has been closed.