You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment, the projected columns are provided just in the API of ConnectorPageSourceProvider.createPageSource(). For the BigQuery connector, the columns needs to be provided at the ConnectorSplitManager as at this point the ReadSession that contains the columns is created.
This issues blocks Issue #2532
This should be possible today with applyProjection(). I ran TpchQueryRunner with a breakpoint in the default implementation in ConnectorMetadata, then ran SELECT count(*) FROM tpch.tiny.region and it's called as expected with an empty projection list.
I don't think this is the direction we want to go. As David said above, the necessary information should be communicated to the connector via the applyProjection API. If that's not happening, we should figure out why.
At the moment, the projected columns are provided just in the API of
ConnectorPageSourceProvider.createPageSource()
. For the BigQuery connector, the columns needs to be provided at theConnectorSplitManager
as at this point theReadSession
that contains the columns is created.This issues blocks Issue #2532
This is per @findepi suggestion at #2532 (comment)
The text was updated successfully, but these errors were encountered: