Re: efficient compare
Date: Sat, 22 Apr 2006 13:27:12 +0200
Message-ID: <444a1311$0$79617$892e7fe2_at_authen.yellow.readfreenews.net>
David Cressey wrote:
>
> I'm not sure what you mean by "efficiently". Could you be a little more
> specific about the envirnment where you intend to do this, and what would
> be wasted by an inefficient solution?
I want to use some kind of hashing to avoid sending the actual data. Inefficient solution would be transferring actual data that is really of no use to the receiver. It is a simple optimization problem, minimize traffic between A and B. Something similar to the RSYNC (by Trigdell) algorithm, but for sets of tuples rather than BLOBs.
> In the meantime, have you tried (A Union B) minus (A intersect B)? Is this
> inefficient? Why?
The two sets are on different computers and initially computer X only knows the set A and computer Y knows set B. So your solution is assuming A and B are knows at a node, which is part of the problem that I want to solve.
> By the way, I'm assuming you don't mean "complement" as in the entire
> universe of discourse minus the intersection.
Right, sorry, Universe=A union B. Received on Sat Apr 22 2006 - 13:27:12 CEST