t+1 - Latest Comments in Python’s hash lookup is insanely good

Re: Python’s hash lookup is insanely good

Name — Sun, 15 Nov 2009 00:24:26 -0000

The lookup is often O(1) but the function that hashes a string so it can be looked up is usually not O(1) and will change depending on how long the string is.

Also, collisions are possible in a hash table and in those cases it will not be O(1) to look those items up.

Re: Python’s hash lookup is insanely good

Matt Wilson — Mon, 02 Nov 2009 16:05:08 -0000

O(1) is a constant amount of time. In the toy hash tables I built in
school, the hashing code took a significant amount of time, even if it
was a fixed amount of time, regardless of the number of elements in my
hash table.

What surprised me with python is how small that time really is. I
wasn't able to separate the cost of hashing an object from a simple
string comparison.

Re: Python’s hash lookup is insanely good

Anonymous — Mon, 02 Nov 2009 15:30:05 -0000

Isn't hash lookup supposed to be an O(1) operation? So why is this so surprising?

http://en.wikipedia.org/wiki/H...

Re: Python’s hash lookup is insanely good

CSMajor — Wed, 01 Jul 2009 21:54:40 -0000

And hash tables and programming in general lol

Re: Python’s hash lookup is insanely good

Matt Wilson — Wed, 08 Apr 2009 22:59:20 -0000

I wish I knew enough about hardware to understand that stuff :)

Matt

Re: Python’s hash lookup is insanely good

NdJ — Wed, 08 Apr 2009 22:54:18 -0000

Not to take away from Python in anyway here but I'd expect the reason you see such amazing performance here has to do with the magic of your CPU cache... none the less it is still impressive !