I am trying to find values below 200 in a 569x462 ndarray I read in from a text file. I want to use numpy.where to find the values... I've tried the following code:
import numpy as np
#Load in text file from Sebago
sebago=np.loadtxt('sebagoWatershedElv.txt',skiprows=6)
#Find where elevation is less than 200
lowElv=np.where(sebago[:]<200)
But the output shows values less than 200, and some values that are greater than 200...
This gives you all values that are less than 200 as one dimensional array:
lowElv = sebago[sebago < 200]
numpy.where returns the indices. So, this sebago[np.where(sebago < 200)] gives the same result as my answer.Minor aside, don't copy (or view) your array [:] unless you're doing something besides 1:1.
import numpy as np
a = np.random.randint(10, size=(8, 8))
# array([[2, 6, 5, 1, 1, 8, 0, 3],
# [4, 7, 5, 4, 9, 6, 1, 8],
# [8, 3, 3, 4, 2, 3, 3, 0],
# [7, 3, 6, 3, 0, 0, 8, 6],
# [5, 7, 7, 0, 7, 4, 8, 6],
# [5, 9, 4, 8, 3, 2, 2, 4],
# [3, 4, 6, 6, 5, 2, 1, 0],
# [3, 7, 6, 4, 4, 4, 1, 3]])
If you feed the Boolean array (e.g. a <= 1) into np.where it returns a list of arrays where the length is equal to the number of dimensions (here, 2). Pass it to array and it turns the coordinate pairs into columns (e.g. (0, 3), (0, 4), which you can see are 0 or 1 in the above random data). Transposing it is handy .T.
np.array(np.where(a <= 1))
# array([[0, 0, 0, 1, 2, 3, 3, 4, 6, 6, 7],
# [3, 4, 6, 6, 7, 4, 5, 3, 6, 7, 6]], dtype=int64)
Just index the array with the Boolean. It flattens it and returns the values.
a[a <= 1]
# array([1, 1, 0, 1, 0, 0, 0, 0, 1, 0, 1])
np.count_nonzero(a <= 1)
# 11
sebago[:] does not perform a copy, it just returns another view onto the values in sebago
np.count_nonzero instead. It is both more readable and considerably faster (x5 on my machine) than using np.sum.