Archives of the TeradataForum

Message Posted: Mon, 23 Jun 2008 @ 17:07:29 GMT

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

Hi,

I am working on multi-value compression of columns for very large tables in our database.

The approach is to prepare new tables with the compressed DDL and thereby load them from the original (uncompressed) tables using select- insert. After several kinds of data validations we are dropping the uncompressed tables after taking their tape backups. The process is going fine and the data in the compressed table is also passing the validation test cases.

What I cannot understand is that when I run collect statistics on the compressed columns for the new and old tables, I am getting different unique value counts for the compressed and uncompressed tables. As per my understanding, compression on a column should not be affecting the uniqueness anyhow.

Can anyone from the forum suggest why this is happening.

Thanks in advance !!

Warm Regards,

Arpan Srivastava

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