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
The operators UnscaledValue and MakeDecimal open up the possibility of doing a lot of aggregations on 64-bit and smaller decimal values without any extra support from cudf.
The main thing we need are JNI changes in cudf to either expose logical_cast with a copy. That is because UnscaledValue is doing a logical cast on a DECIMAL_64 to an INT_64, and MakeDecimal is doing the opposite.
We should also allow decimal in HashAggregateExec and write some tests that verify that this works correctly.
The text was updated successfully, but these errors were encountered:
The operators UnscaledValue and MakeDecimal open up the possibility of doing a lot of aggregations on 64-bit and smaller decimal values without any extra support from cudf.
The main thing we need are JNI changes in cudf to either expose logical_cast with a copy. That is because
UnscaledValue
is doing a logical cast on a DECIMAL_64 to an INT_64, andMakeDecimal
is doing the opposite.We should also allow decimal in HashAggregateExec and write some tests that verify that this works correctly.
The text was updated successfully, but these errors were encountered: