HyukjinKwon f984f6acfe Revert "[SPARK-27870][SQL][PYSPARK] Flush batch timely for pandas UDF (for improving pandas UDFs pipeline)" ... ## What changes were proposed in this pull request? This PR reverts 9c4eb99c52 for the reasons below: 1. An alternative was not considered properly, https://github.com/apache/spark/pull/24734#issuecomment-500101639 https://github.com/apache/spark/pull/24734#issuecomment-500102340 https://github.com/apache/spark/pull/24734#issuecomment-499202982 - I opened a PR https://github.com/apache/spark/pull/24826 2. 9c4eb99c52 fixed timely flushing which behaviour is somewhat hacky and the timing isn't also guaranteed (in case each batch takes longer to process). 3. For pipelining for smaller batches, looks it's better to allow to configure buffer size rather than having another factor to flush ## How was this patch tested? N/A Closes #24827 from HyukjinKwon/revert-flush. Authored-by: HyukjinKwon <gurwls223@apache.org> Signed-off-by: Dongjoon Hyun <dhyun@apple.com>		2019-06-09 08:28:31 -07:00
..
__init__.py	[SPARK-26032][PYTHON] Break large sql/tests.py files into smaller files	2018-11-14 14:51:11 +08:00
mllibutils.py	[SPARK-26033][SPARK-26034][PYTHON][FOLLOW-UP] Small cleanup and deduplication in ml/mllib tests	2018-12-03 14:03:10 -08:00
mlutils.py	[SPARK-26033][PYTHON][TESTS] Break large ml/tests.py file into smaller files	2018-11-18 16:02:15 +08:00
sqlutils.py	[SPARK-27387][PYTHON][TESTS] Replace sqlutils.assertPandasEqual with Pandas assert_frame_equals	2019-04-10 07:50:25 +09:00
streamingutils.py	[SPARK-26856][PYSPARK][FOLLOWUP] Fix UT failure due to wrong patterns for Kinesis assembly	2019-04-02 14:52:56 +09:00
utils.py	Revert "[SPARK-27870][SQL][PYSPARK] Flush batch timely for pandas UDF (for improving pandas UDFs pipeline)"	2019-06-09 08:28:31 -07:00