Details
-
Bug
-
Status: Resolved
-
Minor
-
Resolution: Fixed
-
None
-
None
-
centos 7.3
Description
I have an avro file in hdfs that I am reading in using fastavro, converting to a pandas dataframe and then trying to create an arrow table and get as error:
>>> table=pyarrow.Table.from_pandas(my_dataframe)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "pyarrow/table.pxi", line 746, in pyarrow.lib.Table.from_pandas (/arrow/python/build/temp.linux-x86_64-3.6/lib.cxx:34089)
File "pyarrow/table.pxi", line 346, in pyarrow.lib._dataframe_to_arrays (/arrow/python/build/temp.linux-x86_64-3.6/lib.cxx:30476)
File "pyarrow/array.pxi", line 182, in pyarrow.lib.Array.from_pandas (/arrow/python/build/temp.linux-x86_64-3.6/lib.cxx:22110)
File "pyarrow/error.pxi", line 66, in pyarrow.lib.check_status (/arrow/python/build/temp.linux-x86_64-3.6/lib.cxx:7702)
pyarrow.lib.ArrowNotImplementedError: NotImplemented: null
The avro schema indeed has null fields possible. Is this not implemented? I am using pyarrow 0.5.0. Also, for what I am doing I am not using pandas at all, I just read in the avro and I have a list of dicts and really want to write them to disk in parquet format and am utilizing these steps (which isn't optimal but may be necessary without writing more code of my own).
thanks,
Luke