I ran a report on my prod server (Performance - Top queries by total IO) and found the following query the 2nd highest.
select * from Table1 where Col1 is null order by Col2, Col3, Col4
Table has 100,000 rows. 1 pk and about 30 columns
After running it against engine tuning advisor it advised, if i created a
non clustered index on Col1, Col2, Col3 and Col4 it would increase the performance by 77%.
But when i run the command just by it self it runs and completes instantaneously. What am i missing? Should i create the index to reduce the IO's as the Advisor suggests?
View Complete Post