CODEWITHSUNDEEP
pythonpandasdataframelinear-algebra
william_grisaitis
william_grisaitis

Reputation: 5911

Matrix multiplication of two DataFrames by column name?

I want to compute a row-wise dot product between two DataFrames. In my situation, the column names are the same, but the order might be a little different. (And I want to avoid reordering the columns.)

Is it possible to do a "named matmul" with these two DataFrames?

Specifically I have:

df_1 =
           A      B
index1_0   100    200
index1_1   300    400
index1_2   ...
...

df_2 = 
           B      A
index2_0   0.2    0.8
index2_1   0.6    0.4
index2_2   ...
...

and I wanna get:

           index2_0   index2_1
index1_0   120        160
index1_1   320        360
...

Is there a nice way to do this using the column names?

I'm aware that I can reorder the columns so they line up, and then do matmul with @, but i'm wondering if there's a nicer, more intuitive (readable, maintainable) way.

Upvotes: 1

Views: 320

Answers (1)

user7864386
user7864386

Reputation:

You can first align the DataFrames; then use dot:

out = df1.dot(df2[df1.columns].T)

Output:

          index2_0  index2_1
index1_0     120.0     160.0
index1_1     320.0     360.0

Upvotes: 3

Related Questions