BryanCutler commented on a change in pull request #23900: [SPARK-23836][PYTHON]
Add support for StructType return in Scalar Pandas UDF
URL: https://github.com/apache/spark/pull/23900#discussion_r262330285
##########
File path: python/pyspark/serializers.py
##########
@@ -295,7 +297,34 @@ def create_array(s, t):
raise RuntimeError(error_msg % (s.dtype, t), e)
return array
- arrs = [create_array(s, t) for s, t in series]
+ arrs = []
+ for s, t in series:
+ if t is not None and pa.types.is_struct(t):
+ if not isinstance(s, pd.DataFrame):
+ raise ValueError("A field of type StructType expects a
pandas.DataFrame, "
+ "but got: %s" % str(type(s)))
+
+ # Input partition and result pandas.DataFrame empty, make empty
Arrays with struct
+ if len(s) == 0 and len(s.columns) == 0:
Review comment:
I came across a case where there was an empty partition and when the udf
processed it, the result is an empty Pandas DataFrame with no columns, see
https://github.com/apache/spark/pull/23900/files#diff-d1bd0bd4ceeedd30cc219293a75ad90fR395
I figured it would be pretty confusing for the user to handle these kind of
cases, and it's pretty simple to just check and add an empty struct when this
happens, so that's what this check is for.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
