syntax for numpy masked arrays

masked_numpy_array.py

__author__ = 'Jwely' 

import numpy

# in order to handle Nodata values, we must use floats, not integer datatypes.
array = numpy.array([0, 5, 6, 3, 0, 11, 5, 2, 3, 6, 4, 1, 2, 5, 12], "float32")

# mask this array, where all "0" values are masked
# see also: https://docs.scipy.org/doc/numpy/reference/maskedarray.generic.html
masked_array = numpy.ma.masked_equal(array, 0)

# now lets also mask out the really high values because these are "bad" data
masked_array = numpy.ma.masked_greater(masked_array, 10)

# visual representation of masked array
print("masked array")
print masked_array

# exposes the true nature of the array, values where the mask layer is True are masked.
print("separate data and mask arrays")
print masked_array.data
print masked_array.mask

# we can see that zero values are not included when calculating the minimum
print("minimum value")
print masked_array.min()

# we can see that values above 10 are not included when calculating the maximum
print("maximum value")
print masked_array.max()

# lets find the 75th percentile value
print("75th percentile value")
print numpy.nanpercentile(masked_array.filled(numpy.nan), 75)