Comments for It Depends

Comment on Overheard by Jim Winter

Jim Winter — Wed, 18 May 2011 16:11:24 +0000

Could be worse. He could have said Sybase.

Which only works well if managed by SQL Server.

Comment on This Might Be Useful by dean

dean — Wed, 16 Mar 2011 00:05:10 +0000

That’s a topic big enough for a book.

Basically, you analyze all the missing indexes that look similar and see if you can identify one that would satisfy a number of them. You look at those with only a few accesses with a skeptical eye – odds are high that they are due to ad-hoc querying and not your regular everyday workload, and thus should be ignored.

You look at the queries hitting your database to see if any of the bad ones are the ones causing the engine to think it needs an index on something other than what is already there: a query with an indexed column wrapped in a function (ISNULL, for example) may cause this under some circumstances, and in those cases it is the query that should be rewritten rather than indexes added…

You look at design (if design is under your control) to see if there are any crazy things going on there – did somebody create a table or set of tables with a surrogate key and no alternate key?

I could write a series of posts about what you do if Missing > 0.

Comment on This Might Be Useful by protected static

protected static — Tue, 15 Mar 2011 22:52:46 +0000

So where do you start if Missing > 0?

Comment on I Wonder by dean

dean — Mon, 13 Dec 2010 15:36:15 +0000

Sorry to take so long to approve this!

Anyway, yes, I think so. It would be a tiny, narrow expert system, and I hadn’t thought of it that way.

I intend to try solving this, but there are other SQL projects that demand attention right now.

Comment on I Wonder by rubÃ©n

rubÃ©n — Thu, 09 Dec 2010 21:37:01 +0000

It would be great to figure out how to achieve this using bare SQL queries. Especially the manner in which the information would be organized into tables. It would resemble an expert system, what do you think?

Comment on Procedure Query Plans in Cache by protected static

protected static — Fri, 02 Jul 2010 20:21:17 +0000

Very interesting… I could see that coming in handy.

Comment on I May Be a Day Late, But I Don’t Think I’m a Dollar Short by New Post Up at the SQL Blog « Anacronyms Revisited

New Post Up at the SQL Blog « Anacronyms Revisited — Mon, 12 Apr 2010 14:05:43 +0000

[...] In which Dean enters Celko’s SQL Stumpers contest, but probably posts too late. [...]

Comment on ISNULL vs COALESCE by dean

dean — Fri, 05 Mar 2010 15:11:32 +0000

How bad is it that I am replying to a comment left January 7th? Sigh.

Anyway, perhaps it is. The ISNULL() also exists in Sybase SQL, from which SQL Server was developed. I think it’s an example of the Chicken/Egg Cycle.

Comment on ISNULL vs COALESCE by Graham

Graham — Thu, 07 Jan 2010 19:57:08 +0000

This may be a relic of programming in MS languages. There’s an IsNull() function in every single one.

Comment on ISNULL vs COALESCE by protected static

protected static — Fri, 18 Dec 2009 18:07:14 +0000

Hm. I did not know this, thanks for the tip!