Archives of the TeradataForum

Message Posted: Wed, 16 Jul 2003 @ 15:20:31 GMT

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

Hi, folks,

We've got a philosophical question eating up hundreds of gigs of space on our warehouse. I'm looking for an Occam's Razor to cut this Gordian Knot.

There are some large tables with very fat columns which badly need compressed. I've looked at one that's 250 GB and could be shrunk by ten percent with compression on one column--there are others where a much higher compression ratio could be achieved.

When I've proposed this, I've been told this is a design change which requires a complete round of testing for every application which touches those tables. I'm not seeing this argument--in fact, I so fail to see it that I'm having trouble arguing against it.

It seems like a no-brainer to compress a bloated column. Am I wrong?

Thanks,

John A

P.S. I never did get back to the razor/knot metaphor. Maybe next time.

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