[FIXED] Filter Numpy Array for Values Greater than Preceding Value

Issue

Given the following array how can I filter to produce a new array that contains the values that are less than the next value by at least 3?

In other words, I need to compare each value with its neighbour on the right and add it to a new array if that neighbour is higher by 3 or more.

ex_arr = [1, 2, 3, 8, 9, 10, 12, 16, 17, 23]

desired_arr = [3, 12, 17]

Solution

Use diff to compare the successive values and form a boolean mask, add a False to this array (with np.r_) to account for the last value and perform boolean indexing:

ex_arr = np.array([1, 2, 3, 8, 9, 10, 12, 16, 17, 23])

desired_arr = ex_arr[np.r_[np.diff(ex_arr)>=3, False]]

Or using numpy.nonzero:

desired_arr = ex_arr[np.nonzero(np.diff(ex_arr)>=3)[0]]

Output: array([ 3, 12, 17])

Answered By - mozway