Documentation/behaviour error for selecting by label

[ruidc]

df = pandas.DataFrame([['a'], ['b']], index=[1,2], columns=['x'])
df.loc[[3]] # does not raise KeyError as documented

"ALL of the labels for which you ask, must be in the index or a KeyError will be raised!"
http://pandas.pydata.org/pandas-docs/stable/indexing.html#selection-by-label
and
".loc is strictly label based, will raise KeyError when the items are not found":
http://pandas.pydata.org/pandas-docs/stable/indexing.html#different-choices-for-indexing-loc-iloc-and-ix

also, possibly separate bug:

df.loc[pandas.Index(['a', 'k', 'b']), : ]

DOES fail, so we have an inconsistency between .loc[values,:] and .loc[values]

[jreback]

yeh, let me see if I can fix this inconsistency. indexing code has so many paths.

[jreback]

@ruidc ok, this cleaned up the indexing a bit (well a little).

pls have a look over the PR, the release notes should explain.

[ruidc]

if i understand correctly,

The rule is that at least 1 indexer must be found when using a list-like and .loc
so if I looked for 2 labels, one that was present and one that is not, it would not raise?

---

From: jreback notifications@github.com
To: pydata/pandas pandas@noreply.github.com
Cc: ruidc ruidc@yahoo.com
Sent: Tuesday, 12 August 2014, 17:23
Subject: Re: [pandas] Documentation/behaviour error for selecting by label (#7999)

@ruidc ok, this cleaned up the indexing a bit (well a little).
pls have a look over the PR, the release notes should explain.
—
Reply to this email directly or view it on GitHub.

[jreback]

@ruidc right, you would get a reindex (with the missing) set to np.nan. In realitiy nothing much has changed, fyi, just a slight inconsistency

This is unchanged from 0.14.1 (and from the .ix always worked). It actually is pretty useful
as say you have a big list of labels all their except for 1.

In [1]: df = DataFrame([['a'],['b']],index=[1,2])

In [2]: df.loc[[1,4]]
Out[2]: 
     0
1    a
4  NaN

you can of course

In [3]: df.loc[df.index.isin([1,4])]
Out[3]: 
   0
1  a

[ruidc]

Sure it's primarily that it is not as presently documented

.

[jreback]

@ruidc ahh, can you post what you think should be then?

[ruidc]

"at least one of the labels for which you ask must be in the index..."
instead of

"ALL of the labels for which you ask, must be in the index..."

if that is the correct behavior.

---

From: jreback notifications@github.com
To: pydata/pandas pandas@noreply.github.com
Cc: ruidc ruidc@yahoo.com
Sent: Tuesday, 12 August 2014, 19:29
Subject: Re: [pandas] Documentation/behaviour error for selecting by label (#7999)

@ruidc ahh, can you post what you think should be then?
—
Reply to this email directly or view it on GitHub.

[jreback]

we just had a conv about that, see #8003

[jorisvandenbossche]

@ruidc yes, @jreback, already made a comment about it in the whatsnew in the PR, but as you noted here, the docs should also be updated (as they are wrong now)