[Numpy-discussion] actual, desired versus x, y

Keith Goodman kwgoodman at gmail.com
Thu Jun 3 13:12:26 EDT 2010


On Thu, Jun 3, 2010 at 10:07 AM, Keith Goodman <kwgoodman at gmail.com> wrote:
> Some of the numpy.testing assert functions call the input x and y,
> others call it actual and desired:
>
>>> actual = np.array([1+1j])
>>> desired = np.array([2+2j])
>
>>> assert_almost_equal(actual, desired)
> <snip>
> AssertionError: Items are not equal:
> ACTUAL: [ 1.+1.j]
> DESIRED: [ 2.+2.j]
>
>>> assert_almost_equal(actual.real, desired.real)
> <snip>
> (mismatch 100.0%)
>  x: array([ 1.])
>  y: array([ 2.])
>
> I like the actual and desired, helps me remember which is which.
>
> BTW: numpy.testing is very handy!

I think it would be easier to read the output if the x or actual was
on a separate line. Then the rows of the arrays would line up.

current:

(mismatch 62.5%)
 x: array([[  1.,  NaN,   2.,  NaN,  NaN],
       [  2.,   2.,  NaN,  NaN,  NaN],
       [  3.,   4.,   4.,   1.,  NaN]])
 y: array([[  1.,  NaN,   1.,  NaN,  NaN],
       [  2.,   1.,  NaN,  NaN,  NaN],
       [  3.,   3.,   3.,   2.,  NaN]])

suggested (but alignment will be lost unless you're reading this in a
fix width font):

(mismatch 62.5%)
 x:
     array([[  1.,  NaN,   2.,  NaN,  NaN],
            [  2.,   2.,  NaN,  NaN,  NaN],
            [  3.,   4.,   4.,   1.,  NaN]])
 y:
     array([[  1.,  NaN,   1.,  NaN,  NaN],
            [  2.,   1.,  NaN,  NaN,  NaN],
            [  3.,   3.,   3.,   2.,  NaN]])



More information about the NumPy-Discussion mailing list