LaurentErreca
LaurentErreca
The problem is due to the dtype of the partition column which is set to pd.CategoricalDtype. By converting the dtype to str, the partitioned dataset is correctly read. The user...
Hi, I'm agree with you, it is better to try to support dictionary first, also, the value type in dictionary seems to be well inferred.
What makes me think that pyarrow correctly infer types : When I tried with a column of type object (string values) as partition column, the error contains: ```SchemaError: pyarrow.Field is...
I think that this issue cannot be closed with #306 because we still can't read hive partitioned dataset with native or Dask execution engine.