python 2.7 - parallel indexing in pandas dataframe using a pandas series? -
if have data frame looks this:
a b c d
0 10 15 23 56
1 11 64 66 34
2 46 98 56 87
3 3 49 89 100
4 57 32 48 27
5 99 68 35 19
6 65 79 95 91
8 79 44 66 92
the indicies interested in each column follows (the indicdes in form of series):
a b c d
0 3 5 8 1
so result want is
a b c d
0 3 68 66 34
i want able pull out result
i can convert indicies array if need goal index these different columns in parallel.
you use dataframe.lookup
, , build new frame. not prettiest, but:
>>> s 3 b 5 c 8 d 1 name: 0, dtype: int64 >>> df.lookup(s.values, s.keys()) array([ 3, 68, 66, 34]) >>> pd.dataframe(dict(zip(s.keys(), df.lookup(s.values, s.keys()))), index=[0]) b c d 0 3 68 66 34
where i'm assuming we're using s
, series
, hold indices.
Comments
Post a Comment