mask=[0, 1, 0, 1])
    >>> np.ma.setdiff1d(x, [1, 2])
    masked_array(data=[3, --],
                 mask=[False,  True],
           fill_value=999999)

    TrÞ