Archives of the TeradataForum

Message Posted: Thu, 26 Jun 2003 @ 15:02:59 GMT

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

Joe,

The hashing algorithm is sort of (anyone is welcome to give a more precise description) rolling accumulator of the bytes in the "key" columns, into the 4 byte hash value. As such there is no guarantee that distinct "key" column values won't generate the same hash key. And since the maximum HASHROW value is x"FFFFFFFF" (4+ billion) any table with more than 4+ billion rows would have duplicates.

Your second query says that there are 21 "key" values with more than 1 occurrence, the 1st query - I'm not sure what it is telling me (that there are 3 million rows in the table but only 3M-51607 unique hash values?)

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