Github user ueshin commented on a diff in the pull request:
https://github.com/apache/spark/pull/19246#discussion_r139871986
--- Diff: python/pyspark/sql/types.py ---
@@ -410,6 +410,24 @@ def __init__(self, name, dataType, nullable=True,
metadata=None):
self.dataType = dataType
self.nullable = nullable
self.metadata = metadata or {}
+ self.needConversion = dataType.needConversion
+ self.toInternal = dataType.toInternal
+ self.fromInternal = dataType.fromInternal
+
+ def __getstate__(self):
+ """Return state values to be pickled."""
+ return (self.name, self.dataType, self.nullable, self.metadata)
+
+ def __setstate__(self, state):
+ """Restore state from the unpickled state values."""
+ name, dataType, nullable, metadata = state
+ self.name = name
+ self.dataType = dataType
+ self.nullable = nullable
+ self.metadata = metadata
+ self.needConversion = dataType.needConversion
--- End diff --
What's the difference between your benchmark and @maver1ck's? Why are the
improvements so different?
If the improvement is not quite significant, we shouldn't take this patch
because it confesses developers as you said.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
