How to divide one row by another with same value in a specific column?

Question

I have a dataframe following this pattern:

Measure	Key	mai/21	jun/21	jul/21	…	mar/22	apr/22
Units	4040	100	102	104	…	200	204
Total Value	4040	10500	10700	11000	…	22000	22200
Units	3230	120	130	140	…	230	240
Total Value	3230	120000	130000	140000	…	345000	360000

I need to get a average value per unit and per key, so for every key I have in the dataframe, I need to get the total value row and divide by units row, so I get the average value per unit of each month, and the results needs to follow the same as the input. Just like that:

Measure	Key	mai/21	jun/21	jul/21	…	mar/22	apr/22
Value per Units	4040	105	104.90	105.77	…	110	108.82
Value per Units	3230	1000	1000	1000	…	1500	1500

How can I do it?

Thanks.

Asked By: Lucas Werner

||

Source

Answer 1

You can do:

df2 = df.set_index('Measure').groupby('Key').apply(lambda x: (x.T['Total Value'] / x.T['Units']).drop('Key')).reset_index()
df2['Measure'] = 'Value per units'
df2[df.columns]

For better understanding what lambda does, data for a single key can be extracted and checked separately:

df.set_index('Measure').groupby('Key').get_group(4040)

Besides being more readable, this solution is also about 3x faster than the accepted one.

Answered By: bzu

Answer 2

here is one way to do it, revised after the question was updated

i=0
df2=pd.DataFrame(columns=df.columns)
while (i < len(df)):
    df2.loc[len(df2)] = ['Value per unit', df.iloc[i,1]] + df.iloc[i+1,2:].divide(df.iloc[i,2:] ).tolist()
    i+=2

df2

    Measure          Key    mai/21  jun/21               jul/21     mar/22  apr/22
0   Value per unit  4040    105.0   104.901961       105.769231     110.0   108.823529
1   Value per unit  3230    1000.0  1000.000000     1000.000000     1500.0  1500.000000

Answered By: Naveed

How to divide one row by another with same value in a specific column?

Question:

Answers: