Insert in array at specific location

Question

I have an array [ 0 10 15 20 10 0 35 25 15 35 0 30 20 25 30 0] and I need to insert each element of another array ' [5,7,8,15] ' at locations with an increment of 5 such that the final array looks [ 0 10 15 20 5 10 0 35 25 7 15 35 0 30 8 20 25 30 0 15] length is 20

I am trying with this code

arr_fla = [ 0 10 15 20 10  0 35 25 15 35  0 30 20 25 30  0]
arr_split = [5,7,8,15]
node = 5   
    node_len = node * (node-1)
    
    for w in range(node, node_len, 5):
        for v in arr_split:
            arr_fla = np.insert(arr_fla,w,v)
    print(arr_fla)

The result I am getting is

'[ 0 10 15 20 10 15  8  7  5  0 15  8  7  5 35 15  8  7  5 25 15 35  0 30
 20 25 30  0]' length 28

Can someone please tell me where I am going wrong.

when you insert an element into an array, the elements will shift to the right. Based on your final array, it looks like you want to add only one element from `arr_split` into `arr_fla`. See if your insert statement is doing that — Joe Ferndz, Aug 06 '20 at 11:36
@JoeFerndz In 'arr_fla' I have 16 elements. I want to insert each element of 'arr_split' at positions of 5 , 10 , 15 and 20, so that the array is now of 20 elements. Each element from 'arr_split' exactly once at the respective place. — Gaurav, Aug 06 '20 at 11:43
@JoeFerndz And it is adding all the elements at that position, and I need only one. — Gaurav, Aug 06 '20 at 11:50

score 1 · Answer 1 · answered Aug 06 '20 at 12:00

1

If the sizes line up as cleanly as in your example you can use reshape ...

np.reshape(arr_fla,(len(arr_split),-1))
# array([[ 0, 10, 15, 20],
#        [10,  0, 35, 25],
#        [15, 35,  0, 30],
#        [20, 25, 30,  0]])

... append arr_split as a new column ...

np.c_[np.reshape(arr_fla,(len(arr_split),-1)),arr_split]
# array([[ 0, 10, 15, 20,  5],
#        [10,  0, 35, 25,  7],
#        [15, 35,  0, 30,  8],
#        [20, 25, 30,  0, 15]])

... and flatten again ...

np.c_[np.reshape(arr_fla,(len(arr_split),-1)),arr_split].ravel()
# array([ 0, 10, 15, 20,  5, 10,  0, 35, 25,  7, 15, 35,  0, 30,  8, 20, 25,
#        30,  0, 15])

answered Aug 06 '20 at 12:00

Paul Panzer

51,835
3
54
99

that also seems to give a slight performance plus compared to np.insert – FObersteiner Aug 06 '20 at 12:11
@MrFuppes if you want it fast you should avoid `np.c_` which is for convenience, not speed. And use `np.concatenate` instead. Will have to manually reshape `arr_split`, though. – Paul Panzer Aug 06 '20 at 12:37

score 0 · Accepted Answer · answered Aug 06 '20 at 11:54

I have corrected it:

arr_fla = [0,10,15,20,10,0,35,25,15,35,0,30,20,25,30,0]
arr_split = [5,7,8,15]
node = 5
    
for w in range(len(arr_split)):
    arr_fla = np.insert(arr_fla, (w+1)*node-1, arr_split[w])
print(arr_fla)

'''
Output:
[ 0 10 15 20  5 10  0 35 25  7 15 35  0 30  8 20 25 30  0 15]
'''

In your code:

for v in arr_split:

This gets all the elements at once (in total w times), but you need just one element at a time. Thus you do not need an extra for loop.

score 0 · Answer 3 · answered Aug 06 '20 at 11:55

You want to have a counter that keeps going up every time you insert the item from your second array arr_split.

Try this code. My assumption is that your last element can be inserted directly as the original array has only 16 elements.

arr_fla = [0,10,15,20,10,0,35,25,15,35,0,30,20,25,30,0]
arr_split = [5,7,8,15]

j = 0 #use this as a counter to insert from arr_split

#start iterating from 4th position as you want to insert in the 5th position

for i in range(4,len(arr_fla),5): 

    arr_fla.insert(i,arr_split[j]) #insert at the 5th position every time
    #every time you insert an element, the array size increase

    j +=1 #increase the counter by 1 so you can insert the next element

arr_fla.append(arr_split[j]) #add the final element to the original array 
print(arr_fla)

Output:

[0, 10, 15, 20, 5, 10, 0, 35, 25, 7, 15, 35, 0, 30, 8, 20, 25, 30, 0, 15]

score 0 · Answer 4 · answered Aug 06 '20 at 11:59

You could split the list in even chunks, append to each the split values to each chunk, and reassemble the whole (credit to Ned Batchelder for the chunk function ):

arr_fla = [0,10,15,20,10,0,35,25,15,35,0,30,20,25,30,0]
arr_split = [5,7,8,15]
node = 5  

def chunks(lst, n):
    """Yield successive n-sized chunks from lst."""
    for i in range(0, len(lst), n):
        yield lst[i:i + n]

tmp_arr = chunks(arr_fla, node)

arr_out = []

for index, chunk in enumerate(tmp_arr):
    if arr_split[index]: # make sure arr_split is not exhausted
        chunk.append(arr_split[index]) # we use the index of the chunks list to access the split number to insert 
    arr_out += chunk

print(arr_out)

Outputs:

[0, 10, 15, 20, 10, 5, 0, 35, 25, 15, 35, 7, 0, 30, 20, 25, 30, 8, 0, 15]

score 0 · Answer 5 · answered Aug 06 '20 at 12:12

you can change to below and have a try.

import numpy as np

arr_fla = [0, 10, 15, 20, 10, 0, 35, 25, 15, 35, 0, 30, 20, 25, 30, 0]
arr_split = [5, 7, 8, 15]

index = 4
for ele in arr_split:
    arr_fla = np.insert(arr_fla, index, ele)
    index += 5
print(arr_fla)

the result is

[ 0 10 15 20  5 10  0 35 25  7 15 35  0 30  8 20 25 30  0 15]

about the wrong part of yours, I think it's have two questions:

the second loop is no need, it will cause np insert all the element of arr_split at the same position
the position is not start at 5, it should be 4

Insert in array at specific location

5 Answers5