Replace data in an array with numpy-CodePudding

I am using a digital elevation model as an array, and I want to create a new array with only the pixels below a certain value. I tried using a for loop.

Here is the operation with simpler data :

import numpy as np

array1 = array2 = np.arange(9).reshape(3,3)
array2 = array2.astype("float") #np.nan doesn't work with integers

for row in array2: 
    for item in row:
        if item > 3.0:
            item=np.nan

print(np.array_equal(array1,array2))

The arrays remain equal after the loop is over. Do you know why the values won't be replaced?

CodePudding user response：

You only change item you need insert item in array2, you can use np.where like below:

>>> array1 = np.arange(9).reshape(3,3)
>>> array2 = np.where(array1 > 3, np.nan, array1)
>>> array2
array([[ 0.,  1.,  2.],
       [ 3., nan, nan],
       [nan, nan, nan]])

>>> np.array_equal(array1, array2)
False

CodePudding user response：

Values are not replaced because item is a copy of each array item.

Beside this, using a pure-Python loop is not efficient. Consider vectorizing the code with array2[array2 > 3.0] = np.nan.