Archives of the TeradataForum

Message Posted: Sun, 20 Mar 2005 @ 00:59:48 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hi Geoffrey,

you're right, but even then, what is the chance to update 2 or more rows with the same hashrow value out of 2^32 rows at the same time?

I don't know the hashing algorithm which is used - I think, it quite a good one, but ( as a rule of thumb ) collisions start to occur mainly if more than 80% of the values are already occupied: This means at a 2^32*0.8 level, which is still more than 2^31.

So the chance for 2 rows would be 1:2^31*2^31 --> 1:2^62

I would not bet a cent on this chance ;)

Greetings Christian

	<Prev	Next>		<<First	<Prev	Next>	Last>>