Fixed #12661: more clarification in the where statement

[prabhjotsumman]

• closes DOC: 10 minutes - 'where' not used #12661
• tests added / passed
• passes git diff upstream/master | flake8 --diff
• whatsnew entry

Improvement in the documentation w.r.t to issue #12661

[sinhrks]

Personally, I feel explanation of .where() is inappropriate here. Because the section focuses on df[df.A > 0] and df[df > 0]

Maybe following sentence should be fixed, or pls add an explanation if you add .where() example.

A where operation for getting.

[prabhjotsumman]

ok sir, i will add brief explanation to where

[jorisvandenbossche]

I agree with @sinhrks, I don't think the explanation of where is needed here. It is a section about boolean indexing (which does a 'where' operation), not about the where method. And I don't think it is needed to mention the where method in an 10min intro (there is already a section on where in the docs: http://pandas.pydata.org/pandas-docs/stable/indexing.html#the-where-method-and-masking).

And sorry that this was not really clear from the issue #12661.

I would just remove the ```` around 'where' to make it clear that it is not a method.

[jreback]

So these are not the same at all. I would expand this section and say this is implemented by where internally. Then show an example (and explain the use of where)

In [3]:    dates = pd.date_range('20130101', periods=6)

In [4]:    dates
Out[4]: DatetimeIndex(['2013-01-01', '2013-01-02', '2013-01-03', '2013-01-04', '2013-01-05', '2013-01-06'], dtype='datetime64[ns]', freq='D')

In [5]:    df = pd.DataFrame(np.random.randn(6,4), index=dates, columns=list('ABCD'))

In [6]:    df
Out[6]: 
                   A         B         C         D
2013-01-01 -1.508166 -0.854516  0.148661  1.348457
2013-01-02 -0.890669 -0.329699  0.991305  0.087812
2013-01-03  1.169071 -1.126267 -0.609362 -0.496550
2013-01-04  1.402877 -1.093240  0.038879 -0.042461
2013-01-05 -2.529996  0.570596 -0.556111 -1.365104
2013-01-06  0.036625 -0.241288  0.154433 -1.564450

In [7]: df[df.A>0]
Out[7]: 
                   A         B         C         D
2013-01-03  1.169071 -1.126267 -0.609362 -0.496550
2013-01-04  1.402877 -1.093240  0.038879 -0.042461
2013-01-06  0.036625 -0.241288  0.154433 -1.564450

In [8]: df.where(df>0)
Out[8]: 
                   A         B         C         D
2013-01-01       NaN       NaN  0.148661  1.348457
2013-01-02       NaN       NaN  0.991305  0.087812
2013-01-03  1.169071       NaN       NaN       NaN
2013-01-04  1.402877       NaN  0.038879       NaN
2013-01-05       NaN  0.570596       NaN       NaN
2013-01-06  0.036625       NaN  0.154433       NaN

[jorisvandenbossche]

"this is implemented by where internally" @jreback I don't think this is needed in the 10min section, which we should keep as a simple intro. But it can maybe be added in the boolean indexing section in the indexing docs

[jreback]

ok that sounds fine. this is too much for 10min.

[prabhjotsumman]

I would just remove the `` around 'where' to make it clear that it is not a method.

so, I just remove the `` from where
and also it is given in detail in the indexing section:
http://pandas.pydata.org/pandas-docs/stable/indexing.html#the-where-method-and-masking

[prabhjotsumman]

so, any further work to be done?

[jreback]

closing, but if updated pls reopen