Archives of the TeradataForum

Message Posted: Sat, 09 Jul 2005 @ 05:53:45 GMT

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

Raj:

It sounds like you don't have any true indicators of what constitutes good data. You also do not indicate the volumne of rows nor the number of valid values you anticipate. However, depending on volumes there are two possible solutions for you to persue:

1. put all of the valid value combinations into a temporary or temporal table and join it to your data table to pull only the good ones.

2. write a stored procedure to loop through every column that may have "junk" in it and view each and every character position of the column comparing BETWEEN 'A' and 'Z' OR BETWEEN '0' and '9'. If anything other than these is found, reject it

Hope this helps,

Michael Larkins
Certified Teradata Master
Certified Teradata SQL Instructor

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