calculate mean of columns which only include column value larger than zero

Question

I’d like to calculate the mean of all columns but only consider the columns’ value larger than zero.

For example

df_dict={'A':[0,2,10,0],'B':[10,0,30,40],'C':[0,5,10,10]}
df=pd.DataFrame(df_dict)
df
A   B   C
0   0   10  0
1   2   0   5
2   10  30  10
3   0   40  10

normally, if we just use df.mean(axis=1), it would produce

df.mean(axis=1)
0     3.333333
1     2.333333
2    16.666667
3    16.666667
dtype: float64

what I wanted is: don’t consider any value less than or equal to zero when calculating mean, the result I’d to have is
10/1,(2+5)/2,(10+30+10)/3,(40+10)/2 for each row

How to do it? Thanks

Asked By: roudan

||

Source

Answer 1

First mask the unwanted values. NaNs will be ignored.

With where:

df.where(df.gt(0)).mean(axis=1)

Or with mask:

df.mask(df.le(0)).mean(axis=1)

Output:

0    10.000000
1     3.500000
2    16.666667
3    25.000000
dtype: float64

Answered By: mozway

calculate mean of columns which only include column value larger than zero

Question:

Answers: