Sympy dot product does not read dimensions correctly

Question

I want to multiply two matrices in sympy. However, the .dot() function does not seem to work properly, as when using the transpose matrix so that the dimensions match, the same error appears as when not using it.

Code to reproduce:

from sympy import *
x1, x2, x3 = symbols('x1 x2 x3')

x = Matrix([x1, x2, x3])
m = Matrix([1, 2, 3])
xm = (x-m)
print(xm.shape)
xmt = xm.T
print(xmt.shape)
c = Matrix([[1, 2, 3], [2, 1, 4], [3, 4, 5]])
print(c.shape)

Output:

(3, 1)
(1, 3)
(3, 3)

Now, whether you try

xm.dot(c)

or

xmt.dot(c)

Both commands result in the same error:

ShapeError: Matrix size mismatch: (3, 1) * (3, 3).

which shouldn’t be possible, since xm matrix is indeed of dimensions (3,1) but xmt is of dimensions (1,3).

Am I doing something wrong?

Asked By: liakoyras

||

Source

Answer 1

this seems like a bug in sympy, but using dot is already deprecated, use the @ operator for matrix dot product and it will work

result = xmt @ c

you could post an issue on sympy’s github about dot, but you shouldn’t be using it anyway, as it is deprecated.

Edit: according to @hpaulj the correct function to call in sympy would be xmt.multiply(c) if the @ operator is to be avoided.

Answered By: Ahmed AEK

Sympy dot product does not read dimensions correctly

Question:

Answers: