Invalid syntax using np.zeros [duplicate]

Question

I would like to write this:

x = 'something_{}'.format(1) exec('{} = {}'.format(x,np.zeros((2,2))))

Problem: I get SyntaxError: invalid syntax and I don't know how to solve it.

Someone has an idea?

np.zeros((2,2)) returns an array which you can't use in a format call like this. — elethan
– elethan, Commented Sep 16, 2016 at 16:52
Run "{} = {}".format(x,np.zeros((2,2))) first to see what its value is. You'll see there's an embedded \n, so you'll get an error passing it to exec(). — MattDMo
– MattDMo, Commented Sep 16, 2016 at 18:25

Łukasz Rogalski · Accepted Answer · 2016-09-18 22:31:38Z

String representation of numpy array is not a valid Python literal, therefore it cannot be evaled.

z = np.zeros((2,2)) str(z) # [[ 0. 0.]\n [ 0. 0.]] <-- invalid literal

Technically what you want may be achieved by using repr of object (but in general case it also won't work, e.g. when size of matrix is huge):

import numpy as np x = 'something_{}'.format(1) exec('{} = np.{!r}'.format(x, np.zeros((2,2))))

But what you really want to do is dynamic variable name, and that's a duplicate.

1 Answer 1