Add Trailing Xs Fast

Asked Apr 09 '18 at 23:09

Active Apr 09 '18 at 23:09

Viewed 24 times

I have a list of size N which contains lists of integers (rows) of different lengths which I'd like to convert to a N x M numpy array. I'd like to do this by adding the trailing integer X to the end of every row so that the size of the row becomes M.

The main issue is doing this very quickly.

N ~= 10^6
M ~= 4200

import numpy as np

def addTrailingZeros(data,M,X):
    for i in range(len(data)):
        data[i] = data[i]+[X]*(M-len(data[i]))
    return np.array(data, dtype=int)

data = [[0],[1],[3,3],[8,4,9,3,2,0,5]]
data2 = addTrailingZeros(data,10,0)
print('data: ',data)
print('data2: ',data2)

asked Apr 09 '18 at 23:09

user1581390

1,900
5
25
38

Did you see this question: [Convert and pad a list to numpy array](https://stackoverflow.com/questions/27890052/convert-and-pad-a-list-to-numpy-array)? – miradulo Apr 09 '18 at 23:12
are you trying to add trailing 0s or any int – Primusa Apr 09 '18 at 23:15
@Primusa: 0s right now, but that could change to -1. – user1581390 Apr 09 '18 at 23:29
@miradulo: thanks, I'm looking at it right now, but is it fast? – user1581390 Apr 09 '18 at 23:29
1

@user1581390 I don't know, speed is in the eyes of the benchmarker ;) try it, see if it is fast enough – miradulo Apr 09 '18 at 23:30
@miradulo: that code is more than 10 times slower than what I posted... – user1581390 Apr 09 '18 at 23:55
Because it supports nested lists. I'm assuming you don't have to, then? – miradulo Apr 09 '18 at 23:56
I just have one list of lists, so it is a 2D array in total, not an ND. – user1581390 Apr 09 '18 at 23:59
@user1581390 hpaulj's answer on the linked dup will be fast. – miradulo Apr 10 '18 at 00:03

Add Trailing Xs Fast

0 Answers0