CODEWITHSUNDEEP
pythonpython-2.7numpy
Nikolai
Nikolai

Reputation: 243

nested np.where and boolean array indexing issue

Overview: I am struggling to understand where I am going wrong with this nested mask using np.where. What I am hoping for is- if snow is true, assign 1, if false evaluate the second np.where, which tests if no_snow is true, assign 0, if false (meaning that snow is false and no_snow if false) then assign 2.

# open IMS & pull necessary keys.
hf = h5py.File(ims_dir + 'ims_daily_snow_cover.h5', 'r')
ims = hf['snow_cover'][...]


# create an empty parameter to be later written to new hdf file as gap_fill_flag.
dataset_fill = np.zeros(ims.shape)

# loop through fill - branch based on temporal fill or merra fill.
for day in range(len(fill)):
    # print len(day)
    print day
    fill[day] == 2
    year = days[day][:4]
    # merra fill - more than one consecutive day missing.
    if (fill[day-1] == 2) | (fill[day+1] == 2):
        # run merra_fill function
        # fill with a 2 to signify data are filled from merra.
        ims[day, :] = merra_fill(days[day], ims[day, :])
        dataset_fill[day, :] = 2
    else:
        # temporal_fill - less than one consecutive day missing.
        snow = ((ims[day - 1:day+2, :] == 1).sum(axis=0)) == 2
        no_snow = ((ims[day - 1:day+2, :] == 0).sum(axis=0)) == 2
        # nested np.where.
        ims[day, :] = np.where(snow == True, 1, np.where(no_snow == True, 0, 2))

        dataset_fill[day, :][ims[day, :] < 2] = 1
        dataset_fill[day, :][ims[day, :] == 2] = 2

        ims[day, :][ims[day, :] == 2] = merra_fill(days[day], ims[day, :])

Error:

    ims[day, :] = np.where(snow == True, 1, np.where(no_snow == True, 0, 2))
ValueError: NumPy boolean array indexing assignment cannot assign 2005409 input values to the 0 output values where the mask is true

Help me, stackoverflow. You're my only hope.

Upvotes: 0

Views: 470

Answers (1)

bob.sacamento
bob.sacamento

Reputation: 6651

From help(np.where):

where(condition, [x, y])

Return elements, either from `x` or `y`, depending on `condition`.

If only `condition` is given, return ``condition.nonzero()``.

Parameters
----------
condition : array_like, bool
    When True, yield `x`, otherwise yield `y`.
x, y : array_like, optional
    Values from which to choose. `x` and `y` need to have the same
    shape as `condition`.

I doubt that np.where(no_snow...) has the same shape as snow == True.

Upvotes: 2

Related Questions