Archives of the TeradataForum

Message Posted: Tue, 02 May 2006 @ 09:45:46 GMT

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

     ct T1 (goodpi)

     In TPump,
     using badpi
     ins T1 values(compress(:badpi),....)

     and
     select cols from T1
     where goodpi = ucompress(:badpi)

The idea is to use some algorithm to transform the column content to something that hashes better. It could be a 1 way or 2 way transform. Compression e.g. Huffman could optoinally also reduce the rowsize. Base64 is another 2 way option that should hash better but would increase rather than reduce the rowsize. If round trip were not required e.g. because the badpi value is retained in the row or never rerturned, then round trip would not be required and the algorithm could potentially be checksum, hash, or other 1-way.

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