Divide several columns by several other columns in Pandas

Question

I have a dataset with 20 columns where I want to divide the first 10 by the last 10.

Here are some columns and rows of my dataset:

dict = {'X_1': {'A': 5, 'B': 27, 'C': 26, 'D': 71}, 
        'X_2': {'A': 6, 'B': 51, 'C': 173, 'D': 183},  
        'Y_1': {'A': 479, 'B': 298, 'C': 598, 'D': 693}, 
        'Y_2': {'A': 964, 'B': 474, 'C': 1854, 'D': 1855}
        }

df = pd.DataFrame(dict)

Output:

X_1  X_2  Y_1   Y_2
A    5    6  479   964
B   27   51  298   474
C   26  173  598  1854
D   71  183  693  1855

I could just do it manually,

df["Z_1"] = df["X_1"] / df["Y_1"]
...
df["Z_n"] = df["X_n"] / df["Y_n"]

but I suspect there's a better solution to this.

Any suggestions?

Answer 1

You can do the division in one run using numpy:

df[df.filter(like='X').columns.str.replace('X', 'Z')] = df.filter(like='X').to_numpy() / df.filter(like='Y').to_numpy()

Result:

   X_1  X_2  Y_1   Y_2       Z_1       Z_2
A    5    6  479   964  0.010438  0.006224
B   27   51  298   474  0.090604  0.107595
C   26  173  598  1854  0.043478  0.093312
D   71  183  693  1855  0.102453  0.098652

Answer 2

Use .iloc[] indexing to slice your columns by number...

for i in range(10):
    df.iloc[:,i] / df.iloc[:,10+i]

... then you can form the entire resulting dataframe with a generator expression:

pd.concat((df.iloc[:,i] / df.iloc[:,10+i] for i in range(10)), axis=1)

(There might be a shorter pandas way but I'm not aware of it).

Answer 3

Convert columns to MultiIndex and divide:

temp = df.copy()

temp.columns = temp.columns.str.split("_", expand=True)

temp

    X         Y      
    1    2    1     2
A   5    6  479   964
B  27   51  298   474
C  26  173  598  1854
D  71  183  693  1855

Divide X by Y:

temp = temp.X.div(temp.Y).add_prefix("Z_")

temp
        Z_1       Z_2
A  0.010438  0.006224
B  0.090604  0.107595
C  0.043478  0.093312
D  0.102453  0.098652

Assign new values to original dataframe:

df.assign(**temp)

   X_1  X_2  Y_1   Y_2       Z_1       Z_2
A    5    6  479   964  0.010438  0.006224
B   27   51  298   474  0.090604  0.107595
C   26  173  598  1854  0.043478  0.093312
D   71  183  693  1855  0.102453  0.098652