python 两行相加,Python，Pandas：平均每2行加在一起

pretty basic question, but was wondering:

What is the 'proper' way to average every 2 rows together in pandas Dataframe, and thus end up with only half the number of rows?

Note that this is different than the rolling_mean since it reduces the number of entries.

解决方案

A fast way to do it:

>>> s = pd.Series(range(10))

>>> s

0 0

1 1

2 2

3 3

4 4

5 5

6 6

7 7

8 8

9 9

>>> ((s + s.shift(-1)) / 2)[::2]

0 0.5

2 2.5

4 4.5

6 6.5

8 8.5

The "proper way" I guess would be something like:

>> a = s.index.values

>>> idx = np.array([a, a]).T.flatten()[:len(a)]

>>> idx

[0 0 1 1 2 2 3 3 4 4]

>>> s.groupby(idx).mean()

0 0.5

2 2.5

4 4.5

6 6.5

8 8.5

But it is ~2x slower and gets worse with increasing size.