Re: efficient compare

Bob Badour wrote:
>
> The efficiency will depend on the physical structure representing the
> sets and on the costs of indexes or sorts.

Thats why I am asking, which structure and indexes would be preferable. Assume A and B are on different computers, the goal is to minimize the amount of traffic needed to "unify" these two sets.

> Is this for a course you are taking?

Of course not. I am not looking for a full-fleged solution. I want to know the "state-of-the-art". I assumed the database community would have done lots of work on this. I know a trivial solution, hash each set's key/values and compare, if the differ, split the sets in two pieces (sorted) and exchange... and repeat this until you track the changes. This would require log(n) exchanges to get down to any particular bit. There must be lots of other ways... Received on Fri Apr 21 2006 - 09:44:39 CDT