Curt Sampson wrote:
> On Mon, 16 Feb 2003, Ryan Bradetich wrote:
> > Since my only requirement is that the rows be unique, I have developed a
> > custom MD5 function in C, and created an index on the MD5 hash of the
> > concatanation of all the fields.
>
> Well, that won't guarantee uniqueness, since it's perfectly possible
> to have two different rows hash to the same value. (If that weren't
> possible, your hash would have to contain as much information as the row
> itself, and your space savings wouldn't be nearly so dramatic.)
That's true, but even if he has 4 billion rows it drops the
probability of a duplicate down to something like one in 4 billion, so
it's probably a safe enough bet. His application doesn't require
absolute uniqueness, fortunately, so md5 works well enough in this
case.
Otherwise md5 wouldn't be a terribly good hash...
--
Kevin Brown kevin@sysexperts.com