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
Describe the bug
Currently, cuDF flatten all nested columns before performing the join. After flattening, validity masks of nested type columns are lost. It means that cuDF will regard join key composed by non-null struct who only contains single null child value as a null key by mistake.
For instance,
If the schema of join key is StructType([("a", LongType(nullable=True))], nullable=False), the non-null record Struct([("a", null)]) will be regarded as a null record because of the flattening of nested columns.
The text was updated successfully, but these errors were encountered:
Describe the bug
Currently, cuDF flatten all nested columns before performing the join. After flattening, validity masks of nested type columns are lost. It means that cuDF will regard join key composed by non-null struct who only contains single null child value as a null key by mistake.
For instance,
If the schema of join key is
StructType([("a", LongType(nullable=True))], nullable=False)
, the non-null recordStruct([("a", null)])
will be regarded as a null record because of the flattening of nested columns.The text was updated successfully, but these errors were encountered: