`numpy.lib.scimath:sqrt`

Signature

def sqrt ( x )

Compute the square root of x.

For negative input elements, a complex value is returned (unlike numpy.sqrt which returns NaN).

out : ndarray or scalar: The square root of x. If x was a scalar, so is out, otherwise an array is returned.

For real, non-negative inputs this works just like `numpy.sqrt`:

import numpy as np

✓

np.emath.sqrt(1)
np.emath.sqrt([1, 4])

✗

But it automatically handles negative inputs:

np.emath.sqrt(-1)

✗

np.emath.sqrt([-1,4])

✓

Different results are expected because: floating point 0.0 and -0.0 are distinct. For more control, explicitly use complex() as follows:

np.emath.sqrt(complex(-4.0, 0.0))
np.emath.sqrt(complex(-4.0, -0.0))

✗