对 Pandas 数据框中的特定单元格求和

Question

I have a long Pandas dataframe (10,000 by 5) and I need to take a sum over every 10 cells. My table looks like this.

I was hoping to get my code to look like this, but I'm getting errors.

for i in range(1, 10000):
    if i % 10 == 0:
        avg = df.iloc[i - 10 : i, 3].sum()
        df.iloc[i, 4] = avg

Maybe there's a more Pythonic way to calculate and store averages?

Answer 1

Use GroupBy.transform with sum for new column filled aggregate values, if need fill only last row add mask to DataFrame.loc :

np.random.seed(2020)
df = pd.DataFrame(np.random.randint(10, size=(10000, 4))).add_prefix('col')
  
a = df.index
#if not default index
#a = np.arange(len(df))
df['sum1'] = df.iloc[:, 3].groupby(a // 10).transform('sum')
df.loc[a % 10 == 9, 'sum2'] = df.iloc[:, 3].groupby(a // 10).transform('sum')
print (df.head(20))
   col0  col1  col2  col3  sum1  sum2
0      0     8     3     6    44   NaN
1      3     3     7     8    44   NaN
2      0     0     8     9    44   NaN
3      3     7     2     3    44   NaN
4      6     5     0     4    44   NaN
5      8     6     4     1    44   NaN
6      1     5     9     5    44   NaN
7      6     6     6     5    44   NaN
8      4     6     4     2    44   NaN
9      3     4     7     1    44  44.0
10     4     9     3     2    40   NaN
11     0     9     1     2    40   NaN
12     7     1     0     2    40   NaN
13     8     8     5     6    40   NaN
14     3     3     0     0    40   NaN
15     4     6     6     8    40   NaN
16     9     9     9     5    40   NaN
17     1     9     0     1    40   NaN
18     7     5     0     7    40   NaN
19     1     3     7     7    40  40.0

Answer 2

groupby the row//10 ; take the mean of each group. Does that get you moving?