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
When there is a column in dataframe with all nulls with dtype "object", cudf.read_parquet is inferring the dtype as int32 instead of "object" dtype.
I don't think we want to infer the type to string in this situation (our equivalent of object for now). We could really use an empty or null type in this case.
is it possible NOW to create all-null column with empty type in cudf?
No, we currently don't have a concept of a Null or Empty column in cudf Python currently, but this is something we could plumb in if libcudf supported it.
Describe the bug
When there is a column in dataframe with all nulls with dtype "object",
cudf.read_parquet
is inferring the dtype asint32
instead of"object"
dtype.Steps/Code to reproduce bug
Expected behavior
Expected behavior is to infer the dtype correctly in this scenario.
Environment overview (please complete the following information)
Environment details
Please run and paste the output of the
cudf/print_env.sh
script here, to gather any other relevant environment details(cudf_dev11) pgali@pgali-HP-Z8-G4-Workstation:~/Desktop/cudf$ ./print_env.sh
Additional context
Surfaced while running fuzz tests #6001
The text was updated successfully, but these errors were encountered: