Fastest way to replace values in a numpy array with a list

Question

I want to read a list into a numpy array. This list is being replaced in every iteration of a loop and further operations are done on the array. These operations include element-wise subtraction from another numpy array for a distance measure, and checking a threshold condition in this distance using the numpy.all() function. Currently I am using np.array( list ) each time to convert the list to an array:

#!/usr/bin/python
import numpy as np
a = [1.33,2.555,3.444,5.666,4.555,6.777,8.888]
%timeit b = np.array(a)
100000 loops, best of 3: 4.83 us per loop

Is it possible to do anything better than this, if I know the size of the list and it is invariable? Even small improvements are welcome, as I run this a very large number of times.

I've tried %timeit(np.take(a,range(len(a)),out=b)) which takes much longer: 100000 loops, best of 3: 16.8 us per loop

• plonser: In my opinion there is no faster way of converting a list to a numpy.array. I would rather try to vectorize your for loop. Can you be more specific what you do with the numpy.array ?
• Sahil M: Edited my question with the operations required.
• Warren Weckesser: Where does the list come from? Can you use numpy earlier in your calculations to avoid creating a list?
• Sahil M: @WarrenWeckesser No, I can't use numpy earlier in my calculation, this comes from a simulation engine, which I have little or no control over.
• hpaulj: If the lists aren't that big, it might be faster to do everything with list operations. There is some overhead in creating an array.

Answer 1

As you "know the size of the list and it is invariable", you can set up an array first:

b = np.zeros((7,))

This then works faster:

%timeit b[:] = a
1000000 loops, best of 3: 1.41 μs per loop

vs

%timeit b = np.array(a)
1000000 loops, best of 3: 1.67 μs per loop

• Sahil M: Thanks, this is marginally better. If there is nothing better than this, I'll accept it as an answer,
• Sahil M: I ended up considerably speeding up by using hpaulj's suggestion of not using numpy arrays at all. However this is still the correct and best answer to the above question, thanks!