DataFrame.head(0) doesn't return an empty frame

[trvrm]

pandas.DataFrame({"stuff":[1,2,3,4,5]}).head(0)

returns

   stuff
0      1
1      2
2      3
3      4
4      5

I would expect an empty data frame, in the same way that

[1,2,3,4,5][0:0]

returns an empty list.

This is with pandas version 0.15.0.

The documentation (http://pandas.pydata.org/pandas-docs/version/0.17.0/generated/pandas.DataFrame.head.html) says head(n) returns 'first n rows', so I'd expect 0 rows if n is 0.

[trvrm]

The relevant line is https://github.com/pydata/pandas/blob/49b26600edf4c108a264c7982df8b28a46fac8b1/pandas/core/generic.py#L2135

    def head(self, n=5):
        """
        Returns first n rows
        """
        l = len(self)
        if l == 0 or n==0:
            return self

But I don't see an explanation for this non-intuitive behaviour.

[jreback]

Looks like this was changed as a result of this issue: #5846

but don't really remember why this shouldn't return the same as .iloc[:n]. I suppose its prob more consistent to return the empty.

@TomAugspurger @hayd any thoughts here

[TomAugspurger]

returning an empty frame is probably best.

[jreback]

yep, so prob a propogated 'bug'.

ok, @trvrm want to do a pull-request for this. will be an API change (but tiny).

obviously .tail should be fixed as well.

we will need to confirm tests in .groupby as well (as these are also defined there)

[hayd]

+1

[trvrm]

This pull requests makes head(0) and tail(0) return empty data frames.
#11937

If this is correct, I guess the same needs to be applied to Series.head() and Series.tail()

[hayd]

FYI the Series and groupby version is already correct.