June 5, 2002 at 7:39 pm
When I set statistics profile 'on' in the query window and run a complex query, it gives a lot of information: Rows, executes, statement text, etc, etc. Can you suggest any reading material that explains the various parameters that the statistics profile deals with along with ways to interpret them when analysing queries.
Thanks.
June 5, 2002 at 8:15 pm
Take a look at "SET SHOWPLAN_ALL" as that is the extra fields data you list. There are almost no resources on this on the web and most are rehashes of that data. Sorry, I have not run into any book that goes into too much detail on this. What specifically do you want to know?
"Don't roll your eyes at me. I will tape them in place." (Teacher on Boston Public)
June 6, 2002 at 11:51 am
Here's the list of arguements that I get:
StmtText,StmtId,NodeId,Parent,PhysicalOp,LogicalOp,Argument,DefinedValues,EstimateRows,EstimateIO,EstimateCPU,AvgRowSize,TotalSubtreeCost, OutputList, Warnings Type, Parallel,
EstimateExecutions.
If I am looking to see how efficiently my query is performing, what all should I be looking at. In the physicalop and Logicalop I know that Index seek is what I want rather than Index scan or table scan. In these categories things like 'sort', 'compute scalar', 'aggregate', 'stream aggregate' etc how is that info helping me and what control do I really have on it. Again in the STmtText other than the seek or scan what else do I look for and how do I interpret it.
Guess EstimateRows, EstimateIO, EstimateCPU, AvgRowSize, TotalSubtreeCost should be helpful. Can you explain these somewhat, also are these numbers a real measure or they need to be interpreted in some way. Also, why are all these estimates and not actuals.
Thanks.
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply