Archives of the TeradataForum

Message Posted: Thu, 06 Jan 2011 @ 10:08:16 GMT

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

Adding to following points checking of skewness factor is very imp as well .

Checking for data demographics

Queries for Distribution Demographics

NUmber of Distinct values.

     select count(*) as "total", count(distinct(serial_no)) as "Distinct Values"
     from datatbase.tab1;

* Maximum Rows per value

     SELECT  Serial_no as "Value", COUNT(*) as "Count" FROM  datatbase.tab1 group
     by 1 order by 2 desc;

* Average Rows per value

     SELECT COUNT(*) / COUNT(DISTINCT(Serial_no))  as  "Average Rows Per
     Value"  FROM  datatbase.tab1;

* Space Utilised by Table ampwise.

1. Distinct Values : The More the better

2. Maximum Rows per Value: The fewer the better.

3. Maximum null rows: The fewer the better

4. Typical rows per value: The fewer the better.

Creating join indexes may also help in performance improvement

Regards,

Anshuman Singh
Cognizant (CTS)

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