BUG: reindex strange behaviour

[alefnula]

When I try to reindex a Series object I get a really weird result.

>>> s = pd.Series([1,2,3,4,5], index=['a', 'b', 'c', 'd', 'e'])
>>> s.reindex(['a', 'g', 'c', 'f'], method='ffill')
a     1
g     5
c   NaN
f     5
dtype: float64

According to the documentation I should have got something like:

a   1
g   1
c   3
f   3
dtype: float64

So either the documentation is not valid, or reindex is not working correctly or I haven't understood how this beast works?

[jreback]

is this master or 0.12?

[cpcloud]

Curious. @jreback Should this be a passing test?

(using @alefnula's s)

new_index = list('agcf')
assert_series_equal(s.reindex(new_index, method='ffill'), s.reindex(new_index).ffill())

[cpcloud]

this seems wrong

In [1]: s = Series(range(1,6),index=list(letters[:5]))

In [2]: new_index = list('agcf')

In [3]: s.index.get_indexer(new_index,method='pad')
Out[3]: array([ 0,  4, -1,  4])

[alefnula]

I get the same behaviour on master and on 0.12