I <3 SET STATISTICS IO ON

Once upon a time, a senior DBA at the Salesforce Marketing Cloud had pointed out a very small performance tweak to me. I was curious about whether or not it still held true in SQL Server 2014, so I wrote up the sample below. What’s interesting to me is that a cursory examination of the Execution Plan doesn’t reveal the difference. Of course, if you dig deep enough, you’ll find it there. But it’s not obvious.

On the other hand, by using SET STATISTICS IO ON, I can easily see the difference. Hopefully, this isn’t some new earth-shattering development for you so much as a simple reminder to make use of this statistical information.


IF NOT EXISTS (SELECT 1 FROM sys.indexes WHERE object_id = OBJECT_ID(N'Person.Person') AND name = N'IX_Person_PersonType_EmailPromotion')
CREATE NONCLUSTERED INDEX IX_Person_PersonType_EmailPromotion ON Person.Person (PersonType, EmailPromotion);

SET STATISTICS IO ON; SET NOCOUNT ON;
SELECT BusinessEntityID, EmailPromotion
FROM Person.Person
WHERE PersonType = N'SC'
AND EmailPromotion IN (1, 2);
SELECT BusinessEntityID, EmailPromotion
FROM Person.Person
WHERE PersonType = N'SC'
AND EmailPromotion BETWEEN 1 AND 2;
SET STATISTICS IO OFF;

DROP INDEX IX_Person_PersonType_EmailPromotion ON Person.Person;


Table 'Person'. Scan count 2, logical reads 5, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
Table 'Person'. Scan count 1, logical reads 3, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

Wondering what exactly these results mean? Look here: https://msdn.microsoft.com/en-us/library/ms184361.aspx

Wondering what the tweak was? Look closely at the IN versus the BETWEEN, and then at the number of scans and reads.

Here We Go Again

I’ll be presenting my session from SQL Saturday #358 at this month’s IndyPASS meeting next Tuesday, 3/17. I’m looking forward to all the heckling I’ll be getting from the Salesforce Marketing Cloud‘s performance DBA’s that will be in attendance.

Meeting attendees will also get an added bonus. I have a voucher for a free SQL Admin Toolset from Idera that will be part of the prize drawing. This was courtesy of Idera’s support of #358 in Albuquerque.

So, free pizza, the chance to heckle me, and the chance to win free software… what else could you ask for? See you Tuesday!

[n]varchar truncation

I was very happy with the response to my SQL Saturday session in Albuquerque. To those who attended and provided me with feedback, thank you.

One of the items I covered led to this tweet from @coffegrl:

thrilled! I was just able to use something i learned at #sqlsatabq. that's ROI for you :)

Which just made my day yesterday. Anyway, here’s what she was talking about. Have a look at this script:


DECLARE @stmt NVARCHAR(MAX);
SELECT @stmt = N'foo' + REPLICATE(N'A', 4000) + N'bar';
SELECT LEN(@stmt), LEFT(@stmt, 5), RIGHT(@stmt, 5);

What happens when you run it? Ideally, the results should be 4006, ‘fooAA’ and ‘AAbar’. But that’s not the case. Even though @stmt is NVARCHAR(MAX), the four components of the string concatenation (‘foo’, ‘A’, the results of REPLICATE, and ‘bar’) are not. As a result, the value that is constructed is, in this case, an NVARCHAR(4000), which is then written to @stmt.

Here’s one way to fix it:


DECLARE @stmt NVARCHAR(MAX);
SELECT @stmt = N'foo' + REPLICATE(CAST(N'A' AS NVARCHAR(MAX)), 4000) + N'bar';
SELECT LEN(@stmt), LEFT(@stmt, 5), RIGHT(@stmt, 5);

And another:


DECLARE @stmt NVARCHAR(MAX);
SELECT @stmt = N'foo' + CAST(REPLICATE(N'A', 4000) AS NVARCHAR(MAX)) + N'bar';
SELECT LEN(@stmt), LEFT(@stmt, 5), RIGHT(@stmt, 5);

And another:


DECLARE @stmt NVARCHAR(MAX);
SELECT @stmt = CAST(N'foo' AS NVARCHAR(MAX)) + REPLICATE(N'A', 4000) + N'bar';
SELECT LEN(@stmt), LEFT(@stmt, 5), RIGHT(@stmt, 5);

And the same is true for VARCHAR (assuming you replace the 4000’s with 8000’s). So, the lesson is, when concatenating large strings in SQL be careful that it doesn’t trunca

And So We Meet Again

This is not a technical topic, but one I find myself very passionate about. An article today at NPR brought it to the forefront. As my organization matures, like every organization, rampant movement is slowly replaced by more controlled but slower movement. Notice I used the word “movement”. As Hemingway said, “Never mistake motion for action.” What maturity in an organization often brings is a better ratio of action to motion. We do a better job of working on only those things that really matter, not going off on adventures that may never make their way to production. Or developing with a ready, fire, aim approach. But it also wraps that action in more red tape. It’s a trade-off that every organization encounters.

Anyway, meetings are a highly visible part of that red tape. And the article covers the topic pretty well. I just have one thing to add. It’s a little math that I’ve always kept in the back of my mind at every meeting. Here’s the formula:

Px = T * W

Or, Productivity Lost = Time of the Meeting X the number of Workers at the meeting.

If you are running a meeting for 6 knowledge workers for 2 hours, assume that you just lost 12 hours of productivity. Was that meeting worth those 12 hours? It might have been. But that’s the trade-off. Remember, everyone’s time is valuable.

So, if you have any control over meetings at your organization, please keep that formula in mind. I always do.

Albuquerque SQL Saturday

I’m excited to announce that I’ll be presenting “Fast and Furious Dynamic SQL” at SQL Saturday #358 in Albuquerque on February 7th, 2015. This will be my first time as a PASS presenter, so please save your best heckling and rotten vegetables for the occasion.

The session will cover various tips and tricks that are generally related to running sp_executesql from within a stored procedure. It will focus on performance, security, and maintainability.

Viewing the Plan Cache

I’m working on a script and wanted to know how SQL was caching the query plan. So I wrote this little snippet. The only real value I’m adding here is getting the XQuery correct to pull out the SQL itself.

DBCC FREEPROCCACHE;
-- Run your query here
SELECT
        qp.query_plan,
        qp.query_plan.value(N'(//@StatementText)[1]', N'NVARCHAR(MAX)')
    FROM sys.dm_exec_cached_plans AS cp
        CROSS APPLY sys.dm_exec_query_plan(cp.plan_handle) AS qp;

Of course, I don’t recommend that DBCC in a production environment. But in a dev environment, it makes things easy.

Robinson’s Law

I coined this a while back. I just came across it again, and thought I’d share it. How many of us know this to be true?

“An otherwise inexplicable bug will become easily apparent to the programmer who caused the bug and has subsequently been searching in vain for the bug only and immediately after he announces to his peers that he cannot find the bug, admits defeat, and asks for help.” – Robinson’s Law

Pages Used

One of the items in my toolbox is this simple query that returns the number of pages used by different tables in the database. The nice thing is that it gives one simple number that includes indexes, lob data – basically the whole table.

SELECT TOP 20
    s.name,
    o.name,
    SUM(ddps.used_page_count)
  FROM sys.dm_db_partition_stats AS ddps
    INNER JOIN sys.objects AS o
      ON ddps.object_id = o.object_id
    INNER JOIN sys.schemas AS s
      ON o.schema_id = s.schema_id
  GROUP BY s.name, o.name
  ORDER BY SUM(ddps.used_page_count) DESC;

I just used this on a database today to find that there are some leftover tables from a migration a while back that are eating up a considerable amount of space. Time to run some DROP TABLE scripts!