Getting the dtype of a result array in numpy

Question

I want to preallocate memory for the output of an array operation, and I need to know what dtype to make it. Below I have a function that does what I want it to do, but is terribly ugly.

import numpy as np

def array_operation(arr1, arr2):
    out_shape = arr1.shape
    # Get the dtype of the output, these lines are the ones I want to replace.
    index1 = ([0],) * arr1.ndim
    index2 = ([0],) * arr2.ndim
    tmp_arr = arr1[index1] * arr2[index2]
    out_dtype = tmp_arr.dtype
    # All so I can do the following.
    out_arr = np.empty(out_shape, out_dtype)

The above is pretty ugly. Does numpy have a function that does this?

@Mike Graham: On could enumerate all the basic dtypes, compute the resultant types of all pairs after multiplication, and store result in a dictionary. This wouldn't work for more complicated dtypes though... — unutbu
– unutbu, Commented Sep 2, 2011 at 15:55
@ubutbu, Can you provide an example of a time another operation gives a different resulting original-dtypes-dependent dtype for multiplication and another operation? — Mike Graham
– Mike Graham, Commented Sep 2, 2011 at 16:55
@Mike Graham: I was thinking more about arbitrary (numpy) function calls. — unutbu
– unutbu, Commented Sep 2, 2011 at 17:15

Mike Graham · Accepted Answer · 2011-09-02 15:32:07Z

8

You are looking for numpy.result_type.

(As an aside, do you realize that you can access all multi-dimensional arrays as 1d arrays? You don't need to access x[0, 0, 0, 0, 0] -- you can access x.flat[0].)

edited Sep 2, 2011 at 15:32

answered Sep 2, 2011 at 15:23

Mike Graham

77.2k16 gold badges105 silver badges131 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

kiyo Over a year ago

What numpy version are you using? 1.3.0 doesn't seem to have a result_type function.

kiyo Over a year ago

To your aside: x[0] would 'full slice' all other axes except the first. That could be a lot of extra multiplication.

Mike Graham · Accepted Answer · 2011-09-05 21:38:20Z

1

For those using numpy version < 1.6, you could use:

def result_type(arr1, arr2):
    x1 = arr1.flat[0]
    x2 = arr2.flat[0]
    return (x1 * x2).dtype

def array_operation(arr1, arr2):
    return np.empty(arr1.shape, result_type(arr1, arr2))

This isn't very different from the code you posted, though I think arr1.flat[0] is a slight improvement over index1 = ([0],) * arr1.ndim; arr1[index1].

For numpy version >= 1.6, use Mike Graham's answer, np.result_type

edited Sep 5, 2011 at 21:38

Mike Graham

77.2k16 gold badges105 silver badges131 bronze badges

answered Sep 2, 2011 at 18:27

unutbu

886k197 gold badges1.9k silver badges1.7k bronze badges

Collectives™ on Stack Overflow

Getting the dtype of a result array in numpy

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related