element-wise product matrix x matrix

Question

How to do element-wise product by column, i am looking for a non-loop technique cause i have dozens of columns in each dataframe and different width size

example:

mat1=pd.DataFrame({"a":[1,1,1],"b":[2,2,2]})
mat2=pd.DataFrame({"c":[1,1,1],"d":[1,2,3]})

output

  ac ad bc bd
0 1  1  2  2
1 1  2  2  4
2 1  3  2  6

jezrael · Accepted Answer · 2022-05-23T13:22:35.933

1

Use DataFrame.mul with MultiIndex.from_product and DataFrame.reindex:

mux = pd.MultiIndex.from_product([mat1.columns, mat2.columns])

df = mat1.reindex(mux,  axis=1, level=0).mul(mat2.reindex(mux,  axis=1, level=1))
df.columns=[f'{a}{b}' for a, b in df.columns]
print (df)
   ac  ad  bc  bd
0   1   1   2   2
1   1   2   2   4
2   1   3   2   6

edited May 23 '22 at 13:22

answered May 23 '22 at 12:59

jezrael

822,522
95
1,334
1,252

Sorry, my question wasn't clear at all, i' ve just edited it – Kaoutar May 23 '22 at 13:14
@Kaoutar - Answer was edited. – jezrael May 23 '22 at 13:22
Thank you @jezrael , i've seen this method before, i am wondering if there is some elegant one-line code alternative ? – Kaoutar May 23 '22 at 13:34

score 1 · Answer 2 · answered May 23 '22 at 13:44

1

Easy solution

pd.concat((mat2.mul(mat1[c], axis=0).add_prefix(c) for c in mat1), axis=1)

   ac  ad  bc  bd
0   1   1   2   2
1   1   2   2   4
2   1   3   2   6

answered May 23 '22 at 13:44

Shubham Sharma

68,127
6
24
53

element-wise product matrix x matrix

2 Answers2

Easy solution