Dear all,
reading this :
http://www.brentozar.com/archive/2013/08/query-plans-what-happens-when-row-estimates-get-high/
she said :
"Here’s the general shape of our plan. Just note that the bars connecting operators are narrow– SQL Server has pretty low (and accurate) estimates of how much data will be flowing through these pipes:""
"The plan estimated cost is 1.1 and it estimates it’ll bring back 229 rows. In my test environment it takes 16 ms to run and does several thousand logical reads across the many tables it joins. This means the query optimizer didn’t consider every possible plan, but hey, our runtime was pretty darn good anyway."
how can she know "This means the query optimizer didn’t consider every possible plan, but hey, our runtime was pretty darn good anyway" ?
how can she know "This means the query optimizer didn’t consider every possible plan, but hey, our runtime was pretty darn good anyway" ?
>>
(my) answer to that would be (as well):
16ms (or milliseconds) runtime is pretty darn good no matter what plan and how many logical reads SQL choses - isn't it?
I think thats' what they refer to - I mean is fast regardless SQL consider or not each and any plan and choosing the best - maybe the fastes of all would be 12 (or lets say 9!!) milliseconds comparing to 16ms - would you "feel" that diference? I won't - and pleas don't consider percentage wise but AS IS in milliseconds.