Right, and I don't see any reason for that to decrease just because the
hashtable gets bigger.
A collision is pretty cheap. If the hit ratio for the hash table grows
as it gets bigger, that would account for what you're seeing perhaps.
It's a matter of hash effectiveness, not collision rate.
Henry Rich
On 3/1/2022 4:19 PM, Raul Miller wrote:
Wouldn't 50% full mean approximately a 50% chance of collision with an
arbitrary new value?
Thanks,
--
This email has been checked for viruses by AVG.
https://www.avg.com
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
