Get mean of every 4 rows

user_17003789 0
Viewed 55

I have a df like this:

df = pd.DataFrame({'a': ['2019-09-01 17:00:00', '2019-09-01 17:15:00','2019-09-01 17:30:00','2019-09-01 17:45:00','2019-09-01 18:00:00', '2019-09-01 18:15:00','2019-09-01 18:30:00','2019-09-01 18:45:00'],
                   'b': [432.6, 427.56, 424.2, 433.44,450.24,447.72,452.76,453.6]})

And I want to create a loop to calculate the mean of the values for every 4 items like this: When i = 0 (first position)

mean0 = df.loc[0:3,'b'].mean()

When i = 1:

mean1 = df.loc[4:7,'b'].mean()

And so on.I've tried to create something like this:

for i in df['b]:
    mean[i] = (df[i,'b'] + df.loc[(i+1),'b'] + df.loc[(i+2),'b'])+df.loc[(i+3),'b'])).mean()

But i always get a error message KeyError: 655.7077670000001 or Nan values. Thanks for the help.

pandas dataframe
3 Answers

Try this:

>>> df.groupby(df.index // 4).mean()
        b
0  429.45
1  451.08

Or maybe

>>> df['mean'] = df.groupby(df.index // 4)['b'].transform('mean')
>>> df
                     a       b    mean
0  2019-09-01 17:00:00  432.60  429.45
1  2019-09-01 17:15:00  427.56  429.45
2  2019-09-01 17:30:00  424.20  429.45
3  2019-09-01 17:45:00  433.44  429.45
4  2019-09-01 18:00:00  450.24  451.08
5  2019-09-01 18:15:00  447.72  451.08
6  2019-09-01 18:30:00  452.76  451.08
7  2019-09-01 18:45:00  453.60  451.08
Admin1

Maybe you want to group every 4 values because you want to resample your dataframe.

Try:

out = df.groupby(pd.to_datetime(df['a']).dt.floor('H')).mean().reset_index()
print(out)

# Output
                    a       b
0 2019-09-01 17:00:00  429.45
1 2019-09-01 18:00:00  451.08
user_152399510

This solution is efficient as it is vectorized

mean=list(df.groupby(df.index//4)['b'].mean())

And if you want to continue doing your own method for exploring here is the code

n=df.shape[0]//4
mean=[0]*n
for i in range(n):
  mean[i] = (df.loc[i*4,'b'] + df.loc[(i*4+1),'b'] + df.loc[(i*4+2),'b']+df.loc[(i*4+3),'b'])/4

Output:

[429.45000000000005, 451.08000000000004]

your code was giving you error because .loc was missing here df[i,'b']=>df.loc[i*4,'b']

user_124432600
Related