518 reputation
314
bio website
location
age
visits member for 2 years, 6 months
seen Dec 24 at 18:42

Dec
16
awarded  Self-Learner
Nov
22
awarded  Popular Question
Nov
13
comment How to clear an an Error 15404 after a DB restart (besides rebooting)
Possibly related, we also have an intermittent (weekly) cannot generate SSPI problem on two different servers.
Nov
12
awarded  Popular Question
Nov
7
awarded  Excavator
Nov
7
revised How can I help SQL Server recognize my indexed view column is NOT NULL-able?
Clarify the question headline to "indexed" view to match his example. (Non-indexed view would have a solution.)
Nov
7
comment How can I help SQL Server recognize my indexed view column is NOT NULL-able?
This also appears to be an issue in a recursive cte when recursing on a non-nullable field (no aggregate) although an IsNull(..., 0) at the end can cure.
Nov
7
suggested approved edit on How can I help SQL Server recognize my indexed view column is NOT NULL-able?
Oct
4
comment When should I use a unique constraint instead of a unique index?
Good to see the where clause benefit mentioned.
Aug
23
asked What are recommended (working) permissions for a user to view all SQL Agent jobs
Aug
18
comment Why is an additional filtered statistic being ignored (EAV schema)?
Hmm. If that were true, cases A, C, & D should have no effect but yet they do. Perhaps this is because of the "recompile" option which is mentioned in the cited link and clarified in the comments by Paul White. I will try and make some time to retest to be sure.
Aug
13
comment How can row estimates be improved in order to reduce chances of spills to tempdb
A link to some problematic issues with multi column statistics.
Aug
9
comment Why does Recursive CTE estimate just 1 row?
@ypercube well, it "does not enough math" is a kind of an answer. For instance why not see the top "N" and use "N" as the estimate? A 1 row estimate does not seem probable for any CTE. Since plans may don't change until some cost threshold is crossed maybe they chose 1 since it has same effect as 50 which might be just as bad a guess for a fixed CTE, etc. Along those lines... Is someone certain that anything not based on row statistics will always estimate 1?
Aug
8
comment Why does Recursive CTE estimate just 1 row?
The example is just there to demonstrate the problem. I assume this happens for more general CTE's sometimes (I can't come up with a case right now though). Does no one care that the estimate for a cross of 100 x 100 is 1? The question says "...while keeping the CTE approach." Just answering "Why" for this specific case would be acceptable and help unravel some mystery.
Aug
8
asked Why does Recursive CTE estimate just 1 row?
Jul
25
comment Is support for Parallel Scalar UDF a reasonable feature request?
Clarifying comment. Success with ITVF but not multi-statement TVF.
Jul
24
comment Is support for Parallel Scalar UDF a reasonable feature request?
Sorry, forgot "scalar" in original question, glad you inferred correctly.
Jul
24
accepted Is support for Parallel Scalar UDF a reasonable feature request?
Jul
24
revised Is support for Parallel Scalar UDF a reasonable feature request?
UDF -> Scalar UDF
Jul
24
comment Is support for Parallel Scalar UDF a reasonable feature request?
Yes on success with TVF workaround. I asked because it seems wrong to avoid using such a natural construct. Also it seems impractical to expect new SQL developers to learn UDF internals.