Hyukjin Kwon created SPARK-61172:
------------------------------------
Summary: Add _set_opts to the readers and writers, which
pandas-on-Spark's to_csv and to_json call, in the Python client as PySpark has
it
Key: SPARK-61172
URL: https://issues.apache.org/jira/browse/SPARK-61172
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
PySpark Connect's DataFrameReader, DataFrameWriter, DataFrameWriterV2 and
DataStreamReader have _set_opts(schema=None, **options) of OptionUtils: it sets
the schema, if any, and the options that are not None, through schema and
option. pandas-on-Spark's DataFrame.to_csv and to_json with a path call it on
the writer of the DataFrame (builder._set_opts(sep=..., header=..., ...)).
The Python client's readers and writers have no _set_opts:
AttributeError: 'pyspark.sql.readwriter.DataFrameWriter' object has no
attribute '_set_opts'
Seven tests of Apache Spark's pandas-on-Spark io/test_parity_csv.py and
io/test_parity_dataframe_conversion.py (to_csv and to_json with a path or
partition columns) fail on this client.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
