How to inserts rows in a numpy array based on the condition

Question

I have two lists of numpy arrays and want to insert arrays of a list into arrays of another list:

big_list=[np.array([[1., 1., 2.], [1., 1., 1.9], [1., 1., 5.2]]),
          np.array([[1., 3., 5.1], [0., 0., 1.2], [1., 1., 1.4]])]

small_list= [np.array([[-1., -1., -5.]]),
             np.array([[-1., -1., -1.], [0., -2., -0.5]])]

I want to insert first array of small_list in first array of big_list and also second array of small_list in second array of big_list . I want to put small_list where the third column of big_list has a notable change. In first array of big_list , from second row to third row, the thrid column changes a lot (from 1.9 to 5.2 ). For the second array of big_list , the changes occurs after the first row (from 5.1 to 1.2 ). For example I want to use a threshold like 3 , for the difference of the third column of a row with the next row. So, I want to have:

merged= [np.array([[1., 1., 2.], [1., 1., 1.9], [-1., -1., -5.], [1., 1., 5.2]]),
         np.array([[1., 3., 5.1], [-1., -1., -1.], [0., -2., -0.5], [0., 0., 1.2], [1., 1., 1.4]])]

At the moment, my code can only append rows of small_list at the end of big_list :

merged = [np.append(array, to_append, axis=0) for (array, to_append) in zip(big_list, small_list)]

Is there a way to insert thes rows where I want? In advance, I do appreciate any help.

Answer 1

Try this-

for i, arr in enumerate(big_list):
    diff = list(abs(np.diff(arr[:,-1])) >= 3)
    for t, threshold_state in enumerate(diff):
        if threshold_state:
            big_list[i] = np.insert(big_list[i], t+1, small_list[i], axis=0)

Output

[array([[ 1. ,  1. ,  2. ],
        [ 1. ,  1. ,  1.9],
        [-1. , -1. , -5. ],
        [ 1. ,  1. ,  5.2]]), 
array([[ 1. ,  3. ,  5.1],
        [-1. , -1. , -1. ],
        [ 0. , -2. , -0.5],
        [ 0. ,  0. ,  1.2],
        [ 1. ,  1. ,  1.4]])]

Answer 2

Try this -

• Loop through each element in big_list
• idx calculates the first index where the absolute difference between the last column values is >3
• Insert the corresponding small list in the array at the idx.

merged = []
for i in range(len(big_list)):
    idx = np.argwhere(np.abs(np.diff(big_list[i][:,2]))>3)[0]+1
    m = np.insert(big_list[i], idx, small_list[i], axis=0)
    merged.append(m)
    
merged

[array([[ 1. ,  1. ,  2. ],
        [ 1. ,  1. ,  1.9],
        [-1. , -1. , -5. ],
        [ 1. ,  1. ,  5.2]]),
 array([[ 1. ,  3. ,  5.1],
        [-1. , -1. , -1. ],
        [ 0. , -2. , -0.5],
        [ 0. ,  0. ,  1.2],
        [ 1. ,  1. ,  1.4]])]