[SPARK-11205] [PYSPARK] Match the output of DataFrame#explain() in both scala api and python

[zjffdu]

…rint in python

No test needed. Verify it manually in pyspark shell

[SparkQA]

Test build #43972 has finished for PR 9177 at commit e4417c8.

• This patch fails PySpark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[JoshRosen]

While I agree that it would be nice to have both outputs match, I'm not sure that this is the right way to do it. After your change, the output is being printed by a separate process, breaking the tests and possibly breaking this feature for certain environments where the Py4J Gateway JVM's stdout / stderr isn't captured or display.

If you still want to fix this, can you come up with an approach which prints the header in Python or otherwise harmonizes the output without changing which process performs the printing?

[rxin]

I think this would break a lot of things, including output redirection in Python.

[rxin]

@zjffdu mind closing this pull request?

[zjffdu]

@rxin I will make a new commit soon to match the output in python and scala. Should be a trivial change.

[SparkQA]

Test build #44044 has finished for PR 9177 at commit 39f3fc0.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[rxin]

Can you show the before and after changes?

[JoshRosen]

@rxin, I believe that the doctest itself reflects the before/after.

[rxin]

Thanks - I've merged this.

[zjffdu]

Sorry, @rxin forget to change the title.