Is there a way to tell pandas not to index by character columns? My code is
A=['a','b','c']
B=[1,2,3]
pd.DataFrame(A,B)
0
1 a
2 b
3 c
What I want is just two columns so I can do groupby column A. How do I go about doing this? I can do something like this but I would like to skip the column names to gain as much performance as possible.
pd.DataFrame({'A':A,'B':B})
A B
0 a 1
1 b 2
2 c 3
If you're actually dealing with only two columns, you can group one series by another.
In [6]: A = ['a', 'a', 'b', 'b', 'c', 'c']
In [7]: B = [1, 2, 3, 4, 5, 6]
In [8]: pd.Series(B).groupby(A).mean()
Out[8]:
a 0.5
b 2.5
c 4.5
dtype: float64
I've provided some timings below.
In [9]: %timeit pd.Series(B).groupby(A).mean()
1000 loops, best of 3: 1.07 ms per loop
In [10]: %timeit pd.DataFrame({'A': A, 'B': B}).groupby('A').mean()
100 loops, best of 3: 2.66 ms per loop
In [11]: %timeit pd.DataFrame(list(zip(A, B))).groupby(0).mean()
100 loops, best of 3: 2.38 ms per loop
df = pd.DataFrame([A, B]).Tshould give you what you want.pd.DataFrame(list(zip(A,B)))