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
Currently we have logic for Distinct/GroupBy to "salvage" identifiers, in case they are not part of the distinct projection / grouping key. In case of Distinct we simply can use the entire projection (unless it contains some complex expressions, see #26077), since it's now guaranteed to be unique, and in case of group by we use entire grouping key.
We should also be able to do this for keyless entities - they don't have identifiers so anytime we have a nested collection scenario involving them, we throw due to lack of identifiers needed to perform correct bucketing of the results. For now, we only do this trick if there were any identifiers to begin with.
The text was updated successfully, but these errors were encountered:
Currently we have logic for Distinct/GroupBy to "salvage" identifiers, in case they are not part of the distinct projection / grouping key. In case of Distinct we simply can use the entire projection (unless it contains some complex expressions, see #26077), since it's now guaranteed to be unique, and in case of group by we use entire grouping key.
We should also be able to do this for keyless entities - they don't have identifiers so anytime we have a nested collection scenario involving them, we throw due to lack of identifiers needed to perform correct bucketing of the results. For now, we only do this trick if there were any identifiers to begin with.
The text was updated successfully, but these errors were encountered: