Github user bersprockets commented on the issue:
https://github.com/apache/spark/pull/20909
@HyukjinKwon That makes sense.
Note that when the tests are run using python/run-tests, run-tests.py
steals stdout and stderr. I would need to make a small change to run-tests.py
to detect when a tests is skipped (maybe through retcode) and print the message
(from test test's stdout or stderr).
One other thing. I checked readwriter.py more closely, and there is only a
single docstring test that requires Hive:
<pre>
>>> spark.read.table('tmpTable').dtypes
</pre>
I added <code># doctest: +SKIP</code> to that one line and all the tests
passed. Rather than sometimes skipping all readerwriter tests, maybe we should
just always skip that single test.
udf.py, on the other hand, has lots of docstring tests that require the
test udf files.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
