[Python] Why is this floating point operation faster than integer operation in python 3.11?

So I was solving the 'Median of Two Sorted Arrays' problem on Leetcode and I noticed that the code with integer division somehow slower is than my code with float division then converting the float to integer.

The given cases, the inputs are the following :

• 
  case one : [1,3] and [2]
  
  
• 
  case two : [1,2] and [3,4]
  

So the slower code is the following :

def medianofarrays(array1, array2):
mergedArray = sorted(array1 + array2)
#median is the middle element of the list if the list has uneven amount of elements
if len(mergedArray)%2 != 0:
mean = mergedArray[len(mergedArray)//2]
#median is the mean of the middle 2 elements of the list if the list has an even amount of elements
else:
mean = (mergedArray[len(mergedArray)//2 - 1] + mergedArray[len(mergedArray)//2]) / 2
return mean


And this performed better (time and space) :

def medianofarrays(array1, array2):
mergedArray = sorted(array1 + array2)
#median is the middle element of the list if the list has uneven amount of elements
if len(mergedArray)%2 != 0:
mean = mergedArray[int((len(mergedArray)/2))]
#median is the mean of the middle 2 elements of the list if the list has an even amount of elements
else:
mean = (mergedArray[int((len(mergedArray)/2)-1)] + mergedArray[int(len(mergedArray)/2)]) /2
return mean


Here the time it took

I submitted multiple times to check if it was indeed this case and not some random fluctuation but it gave me the same thing every time. I did not test any further. I expected it to be the other way around, so the int division should be faster right?

Continue reading...