add dropna=False to crosstab example

[testvinder]

>>> foo = pd.Categorical(['a', 'b'], categories=['a', 'b', 'c'])
>>> bar = pd.Categorical(['d', 'e'], categories=['d', 'e', 'f'])
>>> crosstab(foo, bar)  # 'c' and 'f' are not represented in the data,
    ...                     # but they still will be counted in the output
col_0  d  e  f
row_0
a      1  0  0
b      0  1  0
c      0  0  0

The above example code does not produce the output shown because dropna=True is default. Changing crosstab(foo, bar) to crosstab(foo, bar, dropna=False) fixes that and produces the shown output (which is also the expected and correct output).

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

[pep8speaks]

Hello @testvinder! Thanks for updating the PR.

Cheers ! There are no PEP8 issues in this Pull Request. 🍻

Comment last updated on June 11, 2018 at 09:29 Hours UTC

[uds5501]

Please refer to the comments.

@gfyoung , are the suggestions viable?

[uds5501]

@testvinder , just a suggestion, you could've actually just changed the output regarding the given code and mentioned in comments about the default value of dropnabeing set to True

[gfyoung]

I would actually just add an example instead of modifying this one.

[testvinder]

@uds5501 I actually think it's better to set ´dropnatoFalse`. The whole point of specifying the extra categories is to include them in the crosstab. It's a small change but the original is very confusing because the output cannot be reproduced.

[gfyoung]

@testvinder : Ah, yes, your point is well-taken. Can you update the original example with the correct output and add your example with dropna=False ?

[testvinder]

@gfyoung Sure, will do.

[codecov]

Codecov Report

Merging #21413 into master will decrease coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #21413      +/-   ##
==========================================
- Coverage   91.89%   91.89%   -0.01%     
==========================================
  Files         153      153              
  Lines       49596    49596              
==========================================
- Hits        45576    45574       -2     
- Misses       4020     4022       +2

Flag	Coverage Δ
#multiple	90.29% <ø> (-0.01%)	⬇️
#single	41.86% <ø> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/reshape/pivot.py	97.03% <ø> (ø)	⬆️
pandas/util/testing.py	84.6% <0%> (-0.21%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 415012f...4d7e36f. Read the comment docs.

[gfyoung]

I completely misread this PR and thought this was a larger change than it actually was. Sorry!

[testvinder]

@gfyoung yes, it's really just a tiny change that makes the documentation more clear.

[gfyoung]

@testvinder : I just realized: your PR should be against the master branch, not 0.23.x. Could you fix that?

[testvinder]

@gfyoung Done:-) Sorry for the mistake. This is all new to me.

[gfyoung]

@testvinder : I re-applied your commits against master because the base-switching was a lot uglier than I was hoping it would be.

[gfyoung]

With the base fixed, LGTM!

cc @jreback

[jreback]

thanks @testvinder

[jorisvandenbossche]

@jreback just a small note, can you make sure the final commits gets a reasonable message? (normally you should see it in the github merge box), because now it is "Reapply all patches by @testvinder against master" :-)

[jreback]

normally i remove all comments - guess missed that one

[jorisvandenbossche]

Yeah, it's not about the comments (those go into the commit message body) but the title. But it's a bit strange github did it this wrongly, because normally the title is fine (probably because the commit was rebased and squashed)