# How to check numpy arrays are equal

I was doing some exercises in numpy, in particular for broadcasting, but I'm stuck..

Can someone please explain how assert should be used?

```
def fill_0(n):
return np.zeros(n) -1
def fill_1(n):
return np.zeros(n) *(-1)
def fill_2(n):
return - np.ones(n)
def fill_3(n):
return - np.ones(n) -2
assert fill_0(4) == fill_1(4) == fill_2(4) == fill_3(4)
```

Zhang_anlan
commented

So why is it wrong to use np.all() ?

### 1 Answer

I'd do it this way:

```
np.testing.assert_array_equal(fill_0(4), fill_1(4))
np.testing.assert_array_equal(fill_0(4), fill_2(4))
np.testing.assert_array_equal(fill_0(4), fill_3(4))
```

This makes it a lot more clear where the failure is (because each pair is a separate line), and it works even if there are NaNs in the data, whereas regular equality comparison would fail (because NaN==NaN is False).

John Zwinck
posted this

## Have an answer?

JD

I just realized that even the broadcasting is wrong haha that's why I needed to check it