[Numpy-discussion] assigning ma.masked. Strange behavior

[Scott Sinclair]

> 2009/7/4 Ben Park <benparkla at gmail.com>:
>
> import numpy as np
> import numpy.ma as ma
>
> # There is no effect on the following assignment of ma.masked.
> a1 = ma.arange(10).reshape((2,5))
> a1.ravel()[np.array([0,2,2])] = ma.masked

In some situations ravel has to return a copy of the data instead of a
view. You're assigning ma.masked to elements of the copy, not tot
elements of a1.

>>> a1 = ma.arange(10).reshape((2,5))
>>> b = a1.ravel()
>>> b[np.array([0,2,2])] = ma.masked
>>> b
masked_array(data = [-- 1 -- 3 4 5 6 7 8 9],
             mask = [ True False  True False False False False False
False False],
       fill_value = 999999)

>>> a1
masked_array(data =
 [[0 1 2 3 4]
 [5 6 7 8 9]],
             mask =
 False,
       fill_value = 999999)

>>> a1.ravel()[np.array([0,2,2])] = ma.masked
>>> a1
masked_array(data =
 [[0 1 2 3 4]
 [5 6 7 8 9]],
             mask =
 False,
       fill_value = 999999)

Cheers,
Scott