[Mikord]

Originally Posted by Animor Why is it faster to avoid ipairs and simply iterate through the array, as you wrote?

The simplest answer is that ipairs has overhead associated with it since it is an iterator function. The example using the # operator, on the other hand is, just looping from 1 to the number of elements in the array.


Here are some very basic benchmark results between an empty loop iterating over an array with 10,000,000 entries using ipairs versus the # operator. The code is at the bottom of the post:

Generating array with 10000000 entries...
Iterating through 10000000 entries with # operator...
elapsed time: 0.09
Iterating through 10000000 entries with ipairs...
elapsed time: 1.34

local NUM_ITERATIONS = 10000000
 
print(string.format("Generating array with %d entries...", NUM_ITERATIONS))
local array = {}
for i = 1, NUM_ITERATIONS do
    array[i] = i
end
 
print(string.format("Iterating through %d entries with # operator...", #array))
local start = os.clock()
for i = 1, #array do
end
print(string.format("elapsed time: %.2f", os.clock() - start))
 
print(string.format("Iterating through %d entries with ipairs...", #array))
local start = os.clock()
for k, v in ipairs(array) do
end
print(string.format("elapsed time: %.2f", os.clock() - start))