in reply to Re^3: better union of sets algorithm?
in thread better union of sets algorithm?
I'm pretty sure this implies that if you scale the number of buckets proportional to n (instead of n/log(n)), you will get constant-time operations. But perhaps something yucky happens when you try to implement such a creature. Or perhaps I am confusing hashes with other amortized structures (where I know for a fact that certain operations are constant-time). In either case, the empirical evidence suggests I should retract my assertion about constant amortized hashes in Perl. Is there an internals geek around who can shed light on the actual implementation used?
blokhead
|
|---|