Archives of the TeradataForum

Message Posted: Tue, 30 Oct 2007 @ 14:47:00 GMT

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

HI Soumya,

I thought of a query using HashRow and HashBucket functions.

     HASHBUCKET(HASHROW('')) = 0      - Blank Stirng
     HASHBUCKET(HASHROW()) = 65535    - null

Now, Use these combination to get unique id for fd1,fd2,fd3 like this

     Unique_id=HASHBUCKET(HASHROW(fd1))
                   * HASHBUCKET(HASHROW(fd2)) * HASHBUCKET(HASHROW(fd3))

This will give you unique values (also, this is faster as there is no aggrgation is involved) if you do not have values repeated within the fd1,fd2,fd3. That is, It give the same value for (100, 200, 300) , (300,200,100) and (200,300,100).

To get unique in this case also, use any other operation (or combination) other than only multiplication(*)

Hope this gives you hint.

Thanks,

Sridhar Koripally

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