ENH: Explode multiple columns of DataFrame

[kylestahl]

Now .explode() can take a list of column names and will explode multiple at the same time (given that each element across all the columns have the same length in every single row

• closes #xxxx
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[pep8speaks]

Hello @stahl085! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2019-09-16 16:48:53 UTC

[jreback]

always add tests first

[WillAyd]

Just a heads up - this was part of the original implementation in #27267 so you can check there for inspiration on tests and implementation. The big blocker there though was how to handle duplicate values, i.e. whether we should generate a cartesian product or not. Do you know how other similar tools would handle that?

[kylestahl]

Thanks for the info, I like that #27267 implementation much better!
By duplicate do you mean the user passes in the same column name twice? ex: ['A', 'A']

I haven't seen this implemented elsewhere, but to me it seems un-natural for this to return a cartesian product. Would it make sense to include that as an optional argument cartesian=False?

[jreback]

this would need a number of tests; the impl will be very non-performant, so needs updating.