BUG: np.matmul fails on np.matrix of integers

>>> a = np.matrix([[1,1],[1,2]])
>>> a
matrix([[1, 1],
        [1, 2]])
>>> b = np.array([1,2])
>>> b
array([1, 2])
>>> np.matmul(a,b)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
RuntimeError: Iterator automatic output has an array subtype which changed the dimensions of the output
>>> a = a.astype(np.float64)
>>> a
matrix([[ 1.,  1.],
        [ 1.,  2.]])
>>> np.matmul(a,b)
matrix([[ 3.,  5.]])

When a.dtype is np.int32, we cannot do np.matmul on a and b. But after changing the data type to np.float64, it works. Anyone know why is that?

[eric-wieser]

That’s a bizarre error. It might be due to you using np.matrix, which you should really avoid. np.array works just fine for 2d arrays, especially if you’re using matmul

[mattip]

now that matmul is a ufunc, this works

>>> a = np.matrix([[1,1],[1,2]])
>>> b = np.array([1,2])
>>> np.matmul(a,b)
matrix([[3, 5]])