熊猫将帧的移动平均值附加到帧

Question

I can take a SMA of a column and then append it to the dataframe without a problem

BUT then I tried to take a SMA of the column values that met a certain criteria (a subset of all the rows, in my case a sma of goals scored df.f in the last averaging_period homegames df[df.ha=='H'] )

and I get

ValueError: Length of values does not match length of index

My code is wrong, I am taking the moving average of the homegames like I want, but it is removing the row when it gets to an away game, not skipping them in the averaging:

sma = df[df.ha=='H'].f.rolling(window=averaging_period).mean()

df['f_sma%s' % averaging_period] = sma.array

Here are the lengths of the things I am trying to append together

len(df)
   Out[3]: 12938
len(sma.array)
   Out[4]: 6458
len(df[df.ha=='H'])
   Out[5]: 6458

Anyone know how I can get the sma function to not average the goals of the away game, but still retain a row that is the value of the sma up to that point? Therebye returning an array the same length as my dataframe?

Answer 1

You can use .loc filtering when assigning the rolling values to the dataframe

import pandas as pd
import numpy as np

t = pd.DataFrame({'G':['a','a','h','a','h','a','a','h'],'scores':[3,2,1,5,3,1,6,5]})

t['avg'] = np.nan

t.loc[t.G=='h','avg'] = t[t.G=='h'].scores.rolling(window=2).mean()

Output:

   G  scores  avg
0  a       3  NaN
1  a       2  NaN
2  h       1  NaN
3  a       5  NaN
4  h       3  2.0
5  a       1  NaN
6  a       6  NaN
7  h       5  4.0