根据 pandas Dataframe 中的多列替换多列值

Question

A bit of your help will be highly appreciated. I have a dataframe like below:

  name  City kpi       01-Jan-20    02-Jan-20   03-Jan-20   SD       Mean
    1   A    X              9          9    9      0        0        9
    1   A    Y              120       120   120    0        0        120
    2   A    X              10         20   30    10       10        20
    2   A    Y              1          0           11   6.08276253    4
    3   B    X              1          2           3        1         2
    3   B    Y              2          100         5    55.73448962   35.66666667

I want to have an output based on following methods, which takes kpi values for all days(I have 100 days in real df).

def indicator(kpi_value, kpi_std, kpi_mean ):
    SD_distance = abs((kpi_mean-kpi_value)/kpi_std)
    if SD_distance >=1.5:
        return -1
    else:
        return 1

The final output should be like this:

name    City    kpi 01-Jan-20   02-Jan-20   03-Jan-20      SD
1        A       X   1 or -1    1 or -1     1   or -1       0
1        A       Y   1 or -1    1 or -1     1 or -1         0
2        A       X   1 or -1    1 or -1     1 or -1         12.41739671
2        A       Y   1 or -1    1 or -1     1 or -1         0.016969928
3        B       X   1 or -1    1 or -1     1 or -1         0.013699487
3        B       Y   1 or -1    1 or -1     1 or -1         0

I tried with different ways, but could not succeed.

date_cols = df_new.iloc[:,3:-2].columns
df_new[date_cols[0]] = df_new.apply(lambda x: indicator(x[date_cols[0]],x["SD"],x["Mean"])) 

Would be glad to have any suggestions. Thank you!

Answer 1

I think the issue is that you are missing the axis=1 parameter for column-wise in the .apply() function.

date_cols = df_new.iloc[:,3:-2].columns
for col in date_cols:
    df_new[col] = df_new.apply(lambda x: indicator(x[col], x["SD"], x["Mean"]), axis=1)

print(df_new)

Result:

   name city kpi  01-Jan-20  02-Jan-20  03-Jan-20         SD        Mean
0     1    A   X          1          1         -1   0.000000    9.000000
1     1    A   Y          1          1         -1   0.000000  120.000000
2     2    A   X          1          1          1  10.000000   20.000000
3     2    A   Y          1          1          1   6.082763    4.000000
4     3    B   X          1          1          1   1.000000    2.000000
5     3    B   Y          1          1          1  55.734490   35.666667

Note - In your indicator function, the statement SD_distance = abs((kpi_mean-kpi_value)/kpi_std) will give you error if kpi_std is 0. You might want to revisit that logic. Also how did you arrive at the value of SD in your expected result? Is there a logic for it?