Filtered Indexes Work-Around
Filed under: Performance & Tuning, SQL 2008, SQL Tips, Syndication
Recently, I needed to create a stored procedure that queried a rather large table. The table has a filtered index on a date column, and it covers the query. However, the Query Optimizer was not using the index, which was increasing the execution time (not to mention IO!) by at least 10x. This wasn’t the first time I’ve had the Optimizer fail to use a filtered index. Normally when this happens, I use a table hint to force the filtered index — after I verify that it is indeed faster, of course. However, since this was a stored procedure, I was receiving the following error message whenever I tried to execute the proc:
Query processor could not produce a query plan because of the hints defined in this query. Resubmit the query without specifying any hints and without using SET FORCEPLAN.
SQL Server would not allow me to execute the stored procedure using the filtered index hint. If I removed the hint, it executed, but it used a different, non-covering and far more expensive index. For those of you not familiar with this issue, allow me to illustrate the problem.
First, create a table to play with and populate it with some bogus data:
CREATE TABLE dbo.filteredIndexTest ( myID INT IDENTITY(1,3) , myDate SMALLDATETIME , myData CHAR(100) CONSTRAINT PK_filteredIndexTest PRIMARY KEY CLUSTERED(myID) ); SET NOCOUNT ON; DECLARE @DATE SMALLDATETIME = '2010-01-01'; WHILE @DATE < '2010-02-01' BEGIN INSERT INTO dbo.filteredIndexTest ( myDate , myData ) SELECT @DATE , 'Date: ' + CONVERT(VARCHAR(20), @DATE, 102); SET @DATE = DATEADD(MINUTE, 1, @DATE); END; SELECT COUNT(*) FROM dbo.filteredIndexTest;
It looks like this will generate 44,640 rows of test data… plenty enough for our purposes. Now, let’s create our filtered index and write a query that will use it:
CREATE NONCLUSTERED INDEX IX_filteredIndexTest_1 ON dbo.filteredIndexTest(myDate) Include (myData) WHERE myDate >= '2010-01-27'; SELECT DISTINCT myData FROM dbo.filteredIndexTest WHERE myDate >= '2010-01-28';
If you look at the execution plan for this query, you’ll notice that the Optimizer is using the filtered index. Perfect! Now let’s parameterize it.
DECLARE @myDate1 SMALLDATETIME = '2010-01-28'; SELECT DISTINCT myData FROM dbo.filteredIndexTest WHERE myDate >= @myDate1;
Uh oh. Looking at the execution plan, we see that SQL Server is no longer using the filtered index. Instead, it’s scanning the clustered index! Why is this? There’s actually a good explanation for it. The reason is that I could, in theory, pass a date to my parameter that fell outside of the filtered date range. If that’s the case, then SQL Server could not utilize the filtered index. Personally, I think it’s a bug and SQL Server should identify whether or not a filtered index could be used based on the actual value submitted, but… that’s a whole other blog post. 
So what can we do? Well, dynamic SQL may be able to help us out in this case. Let’s give it a go. First, let’s try parameterized dynamic SQL.
DECLARE @mySQL1 NVARCHAR(2000) , @myParam NVARCHAR(2000) = '@p_myDate2 smalldatetime' , @myDate2 SMALLDATETIME = '2010-01-28'; SET @mySQL1 = 'Select Distinct myData From dbo.filteredIndexTest Where myDate >= @p_myDate2'; EXECUTE SP_EXECUTESQL @mySQL1, @myParam, @p_myDate2 = @myDate2;
Looking at the execution plan, we see we’re still scanning on the clustered index. This is because the parameterized dynamic SQL resolves to be the exact same query as the one above it. Let’s try unparameterized SQL instead:
DECLARE @mySQL2 NVARCHAR(2000) , @myDate3 SMALLDATETIME = '2010-01-28'; SET @mySQL2 = 'Select Distinct myData From dbo.filteredIndexTest Where myDate >= ''' + CAST(@myDate3 AS VARCHAR(20)) + ''''; EXECUTE SP_EXECUTESQL @mySQL2; -- Drop Table dbo.filteredIndexTest;
Voila! We have a seek on our filtered index. Why? Because the statement resolves to be identical to our first query, where we hard-coded the date value in the WHERE clause.
Now, I want to stress this fact: you should always, ALWAYS use parameterized dynamic SQL whenever possible. Not only is it safer, but it’s also faster, because it can reuse cached plans. But sometimes you just cannot accomplish the same tasks with it. This is one of those times. If you do end up needing to use unparameterized dynamic SQL as a work-around, please make sure you’re validating your input, especially if you’re interfacing with any sort of external source.
There’s an even easier work-around for this problem that Dave (http://www.crappycoding.com) shared with me: recompile.
Adding “Option (Recompile)” to the end of your statements will force the Optimizer to re-evaluate which index will best meet the needs of your query every time the statement is executed. More importantly, it evaluates the plan based on the actual values passed to the parameter… just like in our hard-coded and dynamic SQL examples. Let’s see it in action:
DECLARE @myDate4 SMALLDATETIME = '2010-01-28'; SELECT DISTINCT myData FROM dbo.filteredIndexTest WHERE myDate >= @myDate4 OPTION (RECOMPILE); DECLARE @myDate5 SMALLDATETIME = '2010-01-20'; SELECT DISTINCT myData FROM dbo.filteredIndexTest WHERE myDate >= @myDate5 OPTION (RECOMPILE);
If we look at the execution plans for the 2 queries above, we see that the first query seeks on the filtered index, and the second query scans on the clustered index. This is because the second query cannot be satisfied with the filtered index because we initially limited our index to dates greater than or equal to 1/27/2010.
There are, of course, trade-offs associated with each approach, so use whichever one best meets your needs. Do you have another work-around for this issue? If so, please let me know.
Update:
Alex Kuznetsov (http://www.simple-talk.com/author/alex-kuznetsov/) shared this method too:
DECLARE @myDate1 SMALLDATETIME = '2010-01-28'; SELECT DISTINCT myData FROM dbo.filteredIndexTest WHERE myDate = @myDate1 AND myDate >= '2010-01-27';
Like the other examples, this will result in an index seek on the filtered index. Basically, by explicitly declaring the start date of your filter, you’re letting the Optimizer know that the filtered index can satisfy the request, regardless of the parameter value passed. Thanks for the tip, Alex!
Index Defrag Script Updates – Beta Testers Needed
Filed under: Miscellaneous, Performance & Tuning, SQL 2008, Syndication, T-SQL Scripts
Update: Wow! I’ve received a ton of responses to my request for beta testers. Thank you all! The SQL Community is really amazing. I’ll hopefully have the new version online in just a few days.
Over the last few months, I’ve received many great comments and suggestions regarding my Index Defrag Script v3.0. I’ve just recently had time to implement most of these suggestions, plus some other things that I thought would be useful.
Here’s some of what you can look forward to shortly:
- Probably the single most requested feature, the new version of the script allows you to set a time limit for index defrags.
- There’s now a static table for managing the status of index defrags. This way, when your time limit is reached, you can pick up where you left off the next day, without the need to rescan indexes.
- There’s now an option to prioritize defrags by range scan counts, fragmentation level, or page counts.
- For those using partitioning, there is now an option to exclude the right-most populated partition from defrags (in theory, the one you’re writing to in a sliding-window scenario).
- Options such as page count limits and SORT_IN_TEMPDB are now parameterized.
- I’ve enhanced error logging.
- … and more!
Right now, I’m looking for a few folks who are willing to beta test the script. If you’re interested, please send me an e-mail at michelle at sqlfool dot com with the editions of SQL Server you can test this on (i.e. 2005 Standard, 2008 Enterprise, etc.).
Thank you!
#PASSAwesomeness
Filed under: Miscellaneous, PASS, Performance & Tuning, Presentations, Syndication
Allen Kinsel on Twitter (@sqlinsaneo) recently started a new Twitter tag, #PASSAwesomeness, about all of the cool things about PASS Summit. I really like the tag, so I’m going to blatantly steal borrow it for this post.
First, and long overdue, I want to give a brief recap of the East Iowa SQL Saturday. On October 17th, our local PASS chapter, 380PASS, sponsored our first ever SQL Saturday at the University of Iowa in Iowa City. By all accounts, the event was a great success! We had 90 attendees, 11 speakers, and 21 sessions. We received numerous compliments on the quality of the speakers, the niceness of the facilities, and the abundance of food. Not too shabby for our first time hosting the event, if I do say so myself.
I’d like to thank all of our wonderful speakers, especially those who traveled from out of town and out of state, for making this event such a success. I’d also like to thank our amazing volunteers for helping put this all together. Lastly, but certainly not least, I’d like to thank our generous sponsors, without whom this event would not be possible. Because this event went so smoothly and was so well received in the community, we’ve already started planning our next big SQL event! In the meantime, don’t forget to check out our monthly 380PASS meetings to tide you over.
I’d also like to take a moment to discuss the PASS Summit. Unless you’re a DBA who’s been living under a rock, you’ve probably heard of the PASS Summit. If you *have* been living under a rock — and hey, I’m not poking fun, I used to live under a rock, too! — then what you need to know is that the Summit is the largest SQL Server conference in the world. It’s a gathering of Microsoft developers and SQL Server gurus; the rest of us show up to try to absorb as much from them as possible. Since I’ve recently moved to the Business Intelligence team, I’m extremely excited to delve into the amazing amount of BI content offered.
I’m also deeply honored to be presenting at the Summit this year on some of the performance tuning techniques I’ve used with great success in my production environments. The session is titled, Super Bowl, Super Load – A Look At Performance Tuning for VLDB’s. If you’re interested in performance tuning or VLDB (very large database) topics, consider stopping by to catch my session. From what I can tell, I’ll be presenting on Tuesday from 10:15am – 11:30am in room(s?) 602-604.
If you read my blog, or if we’ve ever interacted in any way on the internet — Twitter, LinkedIn, e-mails, blog comments, etc. — please stop by and say “hi”! Aside from all of the awesome SQL Server content, I’m really looking forward to meeting as many new folks as possible.
And on that note…
Getting to meet all of the amazing SQL Server professionals out there who have inspired and encouraged me in so many ways #PASSAwesomeness
Partitioning Tricks
Filed under: Miscellaneous, Performance & Tuning, SQL 2008, SQL Tips, Syndication
For those of you who are using partitioning, or who are considering using partitioning, allow me to share some tips with you.
Easy Partition Staging Tables
Switching partitions (or more specifically, hobts) in and out of a partitioned table requires the use of a staging table. The staging table has very specific requirements: it must be completely identical to the partitioned table, including indexing structures, and it must have a check constraint that limits data to the partitioning range. Thanks to my co-worker Jeff, I’ve recently started using the SQL Server Partition Management tool on CodePlex. I haven’t used the automatic partition switching feature — frankly, using any sort of data modification tool in a production environment makes me nervous — but I’ve been using the scripting option to create staging tables in my development environment, which I then copy to production for use. It’s nothing you can’t do yourself, but it does make the whole process easy and painless, plus it saves you from annoying typos. But be careful when using this tool to just create the table and check constraints automatically, because you may need to…
Add Check Constraints After Loading Data
Most of the time, I add the check constraint when I create the staging table, then I load data and perform the partition switch. However, for some reason, I was receiving the following error:
.Net SqlClient Data Provider: Msg 4972, Level 16, State 1, Line 1
ALTER TABLE SWITCH statement failed. Check constraints or partition function of source table ‘myStagingTable’ allows values that are not allowed by check constraints or partition function on target table ‘myDestinationTable’.
This drove me crazy. I confirmed my check constraints were correct, that I had the correct partition number, and that all schema and indexes matched identically. After about 30 minutes of this, I decided to drop and recreate the constraint. For some reason, it fixed the issue. Repeat tests produced the same results: the check constraint needed to be added *after* data was loaded. This error is occurring on a SQL Server 2008 SP1 box; to be honest, I’m not sure what’s causing the error, so if you know, please leave me a comment. But I figured I’d share so that anyone else running into this issue can hopefully save some time and headache.
Replicating Into Partitioned and Non-Partitioned Tables
Recently, we needed to replicate a non-partitioned table to two different destinations. We wanted to use partitioning for Server A, which has 2008 Enterprise; Server B, which is on 2005 Standard, could not take advantage of partitioning. The solution was really easy: create a pre-snapshot and post-snapshot script for the publication, then modify to handle each server group differently. Using pseudo-code, it looked something like this:
/* Identify which servers get the partitioned version */ IF @@SERVERNAME In ('yourServerNameList') BEGIN /* Create your partitioning scheme if necessary */ IF Not Exists(SELECT * FROM sys.partition_schemes WHERE name = 'InsertPartitionScheme') CREATE PARTITION SCHEME InsertPartitionScheme AS PARTITION InsertPartitionFunction ALL TO ([PRIMARY]); /* Create your partitioning function if necessary */ IF Not Exists(SELECT * FROM sys.partition_functions WHERE name = 'InsertPartitionFunction') CREATE PARTITION FUNCTION InsertPartitionFunction (SMALLDATETIME) AS RANGE RIGHT FOR VALUES ('insertValues'); /* Create a partitioned version of your table */ CREATE TABLE [dbo].[yourTableName] ( [yourTableSchema] ) ON InsertPartitionScheme([partitioningKey]); END ELSE BEGIN /* Create a non-partitioned version of your table */ CREATE TABLE [dbo].[yourTableName] ( [yourTableSchema] ) ON [PRIMARY]; END
You could also use an edition check instead of a server name check, if you prefer. The post-snapshot script basically looked the same, except you create partitioned indexes instead.
Compress Old Partitions
Did you know you can set different compression levels for individual partitions? It’s true! I’ve just completed doing this on our largest partitioned table. Here’s how:
/* Apply compression to your partitioned table */ ALTER TABLE dbo.yourTableName Rebuild Partition = All WITH ( Data_Compression = Page ON Partitions(1 TO 9) , Data_Compression = ROW ON Partitions(10 TO 11) , Data_Compression = NONE ON Partitions(12) ); /* Apply compression to your partitioned index */ ALTER INDEX YourPartitionedIndex ON dbo.yourTableName Rebuild Partition = All WITH ( Data_Compression = Page ON Partitions(1 TO 9) , Data_Compression = ROW ON Partitions(10 TO 11) , Data_Compression = NONE ON Partitions(12) ); /* Apply compression to your unpartitioned index */ ALTER INDEX YourUnpartitionedIndex ON dbo.yourTableName Rebuild WITH (Data_Compression = ROW);
A couple of things to note. In all of our proof-of-concept testing, we found that compression significantly reduced query execution time, reads (IO), and storage. However, CPU was also increased significantly. The results were more dramatic, both good and bad, with page compression versus row compression. Still, for our older partitions, which aren’t queried regularly, it made sense to turn on page compression. The newer partitions receive row compression, and the newest partitions, which are still queried very regularly by routine processes, were left completely uncompressed. This seems to strike a nice balance in our environment, but of course, results will vary depending on how you use your data.
Something to be aware of is that compressing your clustered index does *not* compress your non-clustered indexes; those are separate operations. Lastly, for those who are curious, it took us about 1 minute to apply row compression and about 7 minutes to apply page compression to partitions averaging 30 million rows.
Looking for more information on table partitioning? Check out my overview of partitioning, my example code, and my article on indexing on partitioned tables.
Monitoring Process for Performance Counters
Filed under: Performance & Tuning, Syndication, T-SQL Scripts
Recently I needed to create a process to monitor performance counters over a short period of time. We were going to implement a change and we wanted to compare performance before and after to see if there was any impact.
To do this, I first created a couple of tables. One table is used to actually store the monitored values. The second table is used for configuration; you insert only the counters you want to monitor.
/* Create the table to store our logged perfmon counters */ CREATE TABLE dbo.dba_perfCounterMonitor ( capture_id INT IDENTITY(1,1) Not Null , captureDate SMALLDATETIME Not Null , objectName NVARCHAR(128) Not Null , counterName NVARCHAR(128) Not Null , instanceName NVARCHAR(128) Not Null , VALUE FLOAT(6) Not Null , valueType NVARCHAR(10) Not Null CONSTRAINT PK_dba_perfCounterMonitor PRIMARY KEY CLUSTERED(capture_id) ); /* Create the table that controls which counters we're going to monitor */ CREATE TABLE dbo.dba_perfCounterMonitorConfig ( objectName NVARCHAR(128) Not Null , counterName NVARCHAR(128) Not Null , instanceName NVARCHAR(128) Null );
If you leave the instanceName NULL in the config table, it’ll monitor all instances. Now we’re going to insert some sample performance counters into the config table. The counters you’re interested in can, and likely will, vary.
/* Insert some perfmon counters to be monitored */ INSERT INTO dbo.dba_perfCounterMonitorConfig SELECT 'SQLServer:Buffer Manager', 'Page Life Expectancy', Null UNION All SELECT 'SQLServer:Locks', 'Lock Requests/sec', Null UNION All SELECT 'SQLServer:Locks', 'Lock Waits/sec', Null UNION All SELECT 'SQLServer:Locks', 'Lock Wait Time (ms)', Null UNION All SELECT 'SQLServer:Buffer Manager', 'Page reads/sec', Null UNION All SELECT 'SQLServer:Buffer Manager', 'Page writes/sec', Null UNION All SELECT 'SQLServer:Buffer Manager', 'Buffer cache hit ratio', Null UNION All SELECT 'SQLServer:Databases', 'Transactions/sec', 'AdventureWorks' UNION All SELECT 'SQLServer:General Statistics', 'Processes blocked', Null;
Now let’s create our proc. This proc will run for a specified time period and will *average* the counters over that time. I personally take snapshots every 15 seconds for 4 minutes; I have a scheduled task that runs this every 5 minutes. It’s not perfect, but it gives me a good idea of what’s happening on the server.
CREATE PROCEDURE dbo.dba_perfCounterMonitor_sp /* Declare Parameters */ @samplePeriod INT = 240 /* how long to sample, in seconds */ , @sampleRate CHAR(8) = '00:00:15' /* how frequently to sample, in seconds */ , @displayResults BIT = 0 /* display the results when done */ AS /********************************************************************************* Name: dba_perfCounterMonitor_sp Author: Michelle Ufford, http://sqlfool.com Purpose: Monitors performance counters. Uses the dba_perfCounterMonitorConfig table to manage which perf counters to monitor. @samplePeriod - specifies how long the process will try to monitor performance counters; in seconds. @sampleRate - how long inbetween samples; in seconds. The average values over sample period is then logged to the dba_perfCounterMonitor table. Notes: There are 3 basic types of performance counter calculations: Value/Base: these calculations require 2 counters. The value counter (cntr_type = 537003264) has to be divided by the base counter (cntr_type = 1073939712). Per Second: these counters are store cumulative values; the value must be compared at 2 different times to calculate the difference (cntr_type = 537003264). Point In Time: these counters show what the value of the counter is at the current point-in-time (cntr_type = 65792). No calculation is necessary to derive the value. Called by: DBA Date User Description ---------------------------------------------------------------------------- 2009-09-04 MFU Initial Release ********************************************************************************* Exec dbo.dba_perfCounterMonitor_sp @samplePeriod = 60 , @sampleRate = '00:00:01' , @displayResults = 1; *********************************************************************************/ SET NOCOUNT ON; SET XACT_Abort ON; SET Ansi_Padding ON; SET Ansi_Warnings ON; SET ArithAbort ON; SET Concat_Null_Yields_Null ON; SET Numeric_RoundAbort OFF; BEGIN /* Declare Variables */ DECLARE @startTime DATETIME , @endTime DATETIME , @iteration INT; SELECT @startTime = GETDATE() , @iteration = 1; DECLARE @samples TABLE ( iteration INT Not Null , objectName NVARCHAR(128) Not Null , counterName NVARCHAR(128) Not Null , instanceName NVARCHAR(128) Not Null , cntr_value FLOAT Not Null , base_value FLOAT Null , cntr_type BIGINT Not Null ); BEGIN Try /* Start a new transaction */ BEGIN TRANSACTION; /* Grab all of our counters */ INSERT INTO @samples SELECT @iteration , RTRIM(dopc.OBJECT_NAME) , RTRIM(dopc.counter_name) , RTRIM(dopc.instance_name) , RTRIM(dopc.cntr_value) , (SELECT cntr_value FROM sys.dm_os_performance_counters AS dopc1 WHERE dopc1.OBJECT_NAME = pcml.objectName And dopc1.counter_name = pcml.counterName + ' base' And dopc1.instance_name = IsNull(pcml.instanceName, dopc.instance_name)) , dopc.cntr_type FROM sys.dm_os_performance_counters AS dopc Join dbo.dba_perfCounterMonitorConfig AS pcml ON dopc.OBJECT_NAME = pcml.objectName And dopc.counter_name = pcml.counterName And dopc.instance_name = IsNull(pcml.instanceName, dopc.instance_name); /* During our sample period, grab our counter values and store the results */ WHILE GETDATE() < DATEADD(SECOND, @samplePeriod, @startTime) BEGIN SET @iteration = @iteration + 1; INSERT INTO @samples SELECT @iteration , RTRIM(dopc.OBJECT_NAME) , RTRIM(dopc.counter_name) , RTRIM(dopc.instance_name) , dopc.cntr_value , (SELECT cntr_value FROM sys.dm_os_performance_counters AS dopc1 WHERE dopc1.OBJECT_NAME = pcml.objectName And dopc1.counter_name = pcml.counterName + ' base' And dopc1.instance_name = IsNull(pcml.instanceName, dopc.instance_name)) , dopc.cntr_type FROM sys.dm_os_performance_counters AS dopc Join dbo.dba_perfCounterMonitorConfig AS pcml ON dopc.OBJECT_NAME = pcml.objectName And dopc.counter_name = pcml.counterName And dopc.instance_name = IsNull(pcml.instanceName, dopc.instance_name); /* Wait for a small delay */ WAITFOR Delay @sampleRate; END; /* Grab our end time for calculations */ SET @endTime = GETDATE(); /* Store the average of our point-in-time counters */ INSERT INTO dbo.dba_perfCounterMonitor ( captureDate , objectName , counterName , instanceName , VALUE , valueType ) SELECT @startTime , objectName , counterName , instanceName , AVG(cntr_value) , 'value' FROM @samples WHERE cntr_type = 65792 GROUP BY objectName , counterName , instanceName; /* Store the average of the value vs the base for cntr_type = 537003264 */ INSERT INTO dbo.dba_perfCounterMonitor ( captureDate , objectName , counterName , instanceName , VALUE , valueType ) SELECT @startTime , objectName , counterName , instanceName , AVG(cntr_value)/AVG(IsNull(base_value, 1)) , 'percent' FROM @samples WHERE cntr_type = 537003264 GROUP BY objectName , counterName , instanceName; /* Compare the first and last values for our cumulative, per-second counters */ INSERT INTO dbo.dba_perfCounterMonitor ( captureDate , objectName , counterName , instanceName , VALUE , valueType ) SELECT @startTime , objectName , counterName , instanceName , (MAX(cntr_value) - MIN(cntr_value)) / DATEDIFF(SECOND, @startTime, @endTime) , 'value' FROM @samples WHERE cntr_type = 272696576 GROUP BY objectName , counterName , instanceName; /* Should we display the results of our most recent execution? */ IF @displayResults = 1 SELECT captureDate , objectName , counterName , instanceName , VALUE , valueType FROM dbo.dba_perfCounterMonitor WITH (NoLock) WHERE captureDate = CAST(@startTime AS SMALLDATETIME) ORDER BY objectName , counterName , instanceName; /* If you have an open transaction, commit it */ IF @@TRANCOUNT > 0 COMMIT TRANSACTION; END Try BEGIN Catch /* Whoops, there was an error... rollback! */ IF @@TRANCOUNT > 0 ROLLBACK TRANSACTION; /* Return an error message and log it */ EXECUTE dbo.dba_logError_sp; END Catch; SET NOCOUNT OFF; RETURN 0; END Go
Like I said, it’s not perfect, but it gets the job done.
Getting an error about dba_logError_sp? Take a look at my error handling proc.
Index Defrag Script, v3.0
Filed under: Performance & Tuning, SQL 2008, Syndication, T-SQL Scripts
I’ve just completed the latest version of my index defrag script! Here’s a brief list of the updates:
- Fixed a bug with the LOB logic. In the previous version, after a LOB was encountered, all subsequent indexes would be reorganized.
- Added support for stat rebuilds after the index defrag is complete (@rebuildStats)
- Added an exclusion list table (dba_indexDefragExclusion) to support index scheduling
- Modified logging to show which defrags are “in progress”; added columns to dba_indexDefragLog
- Added support for the defrag of the model and msdb databases
- Added @scanMode as a configurable parameter
So what can this index defrag script do? Well, for starters, you can:
- Schedule it to run with the default settings; it works “right out of the box” with no additional configuration necessary
- Run this one script from a centralized database for all databases on a server
- Run this script for a specific database or table
- Configure custom threshold limits and the point at which a rebuild should be performed (instead of a reorganize)
- Defrag individual partitions
- Log its actions and the duration of the defrag
- Run in “commands only” mode (@executeSQL = 0, @printCommands = 1)
- Customize performance parameters such as @maxDopRestriction and @defragDelay to minimize impact on the server
- Schedule specific indexes to only be defragged on weekends, or every other day
To use this last option, you need to add a record to the dba_indexDefragExclusion table. I think all of the columns are pretty self-explanatory except the [exclusionMask] column. The way this works is each day of the week is assigned a value:
1=Sunday, 2=Monday, 4=Tuesday, 8=Wednesday, 16=Thursday, 32=Friday, 64=Saturday
Take a SUM of the values for the days that you want excluded. So if you want an index to only be defragged on weekends, you would add up Monday through Friday (2+4+8+16+32) and use a value of 62 for the exclusionMask column. For a little more information on how this works, check out my blog post on Bitwise Operations.
Please note: if you don’t insert any records into the dba_indexDefragExclusion table, by default all indexes will be defragged every run-time if they exceed the specified thresholds. This is normal behavior and may be perfectly fine in your environment. However, if the dba_indexDefragExclusion table does not exist, the script will fail.
I try to document each parameter within the code, so check the comments section in the script for a full list of parameters and what they do.
Special thanks to everyone who helped beta test this script!
Without further ado, the script:
/* Drop Table Scripts: Drop Table dbo.dba_indexDefragLog; Drop Table dbo.dba_indexDefragExclusion; */ IF Not Exists(SELECT [OBJECT_ID] FROM sys.tables WHERE [name] In (N'dba_indexDefragLog', 'dba_indexDefragExclusion')) BEGIN CREATE TABLE dbo.dba_indexDefragLog ( indexDefrag_id INT IDENTITY(1,1) Not Null , databaseID INT Not Null , databaseName NVARCHAR(128) Not Null , objectID INT Not Null , objectName NVARCHAR(128) Not Null , indexID INT Not Null , indexName NVARCHAR(128) Not Null , partitionNumber SMALLINT Not Null , fragmentation FLOAT Not Null , page_count INT Not Null , dateTimeStart DATETIME Not Null , dateTimeEnd DATETIME Null , durationSeconds INT Null CONSTRAINT PK_indexDefragLog PRIMARY KEY CLUSTERED (indexDefrag_id) ); PRINT 'dba_indexDefragLog Table Created'; CREATE TABLE dbo.dba_indexDefragExclusion ( databaseID INT Not Null , databaseName NVARCHAR(128) Not Null , objectID INT Not Null , objectName NVARCHAR(128) Not Null , indexID INT Not Null , indexName NVARCHAR(128) Not Null , exclusionMask INT Not Null /* 1=Sunday, 2=Monday, 4=Tuesday, 8=Wednesday, 16=Thursday, 32=Friday, 64=Saturday */ CONSTRAINT PK_indexDefragExclusion PRIMARY KEY CLUSTERED (databaseID, objectID, indexID) ); PRINT 'dba_indexDefragExclusion Table Created'; END ELSE RAISERROR('One or more tables already exist. Please drop or rename before proceeding.', 16, 0); IF OBJECTPROPERTY(OBJECT_ID('dbo.dba_indexDefrag_sp'), N'IsProcedure') = 1 BEGIN DROP PROCEDURE dbo.dba_indexDefrag_sp; PRINT 'Procedure dba_indexDefrag_sp dropped'; END; Go CREATE PROCEDURE dbo.dba_indexDefrag_sp /* Declare Parameters */ @minFragmentation FLOAT = 5.0 /* in percent, will not defrag if fragmentation less than specified */ , @rebuildThreshold FLOAT = 30.0 /* in percent, greater than @rebuildThreshold will result in rebuild instead of reorg */ , @executeSQL BIT = 1 /* 1 = execute; 0 = print command only */ , @DATABASE VARCHAR(128) = Null /* Option to specify a database name; null will return all */ , @tableName VARCHAR(4000) = Null -- databaseName.schema.tableName /* Option to specify a table name; null will return all */ , @scanMode VARCHAR(10) = N'LIMITED' /* Options are LIMITED, SAMPLED, and DETAILED */ , @onlineRebuild BIT = 1 /* 1 = online rebuild; 0 = offline rebuild; only in Enterprise */ , @maxDopRestriction TINYINT = Null /* Option to restrict the number of processors for the operation; only in Enterprise */ , @printCommands BIT = 0 /* 1 = print commands; 0 = do not print commands */ , @printFragmentation BIT = 0 /* 1 = print fragmentation prior to defrag; 0 = do not print */ , @defragDelay CHAR(8) = '00:00:05' /* time to wait between defrag commands */ , @debugMode BIT = 0 /* display some useful comments to help determine if/where issues occur */ , @rebuildStats BIT = 1 /* option to rebuild stats after completed index defrags */ AS /********************************************************************************* Name: dba_indexDefrag_sp Author: Michelle Ufford, http://sqlfool.com Purpose: Defrags all indexes for the current database Notes: CAUTION: TRANSACTION LOG SIZE SHOULD BE MONITORED CLOSELY WHEN DEFRAGMENTING. @minFragmentation defaulted to 10%, will not defrag if fragmentation is less than that @rebuildThreshold defaulted to 30% as recommended by Microsoft in BOL; greater than 30% will result in rebuild instead @executeSQL 1 = execute the SQL generated by this proc; 0 = print command only @database Optional, specify specific database name to defrag; If not specified, all non-system databases will be defragged. @tableName Specify if you only want to defrag indexes for a specific table, format = databaseName.schema.tableName; if not specified, all tables will be defragged. @scanMode Specifies which scan mode to use to determine fragmentation levels. Options are: LIMITED - scans the parent level; quickest mode, recommended for most cases. SAMPLED - samples 1% of all data pages; if less than 10k pages, performs a DETAILED scan. DETAILED - scans all data pages. Use great care with this mode, as it can cause performance issues. @onlineRebuild 1 = online rebuild; 0 = offline rebuild @maxDopRestriction Option to specify a processor limit for index rebuilds @printCommands 1 = print commands to screen; 0 = do not print commands @printFragmentation 1 = print fragmentation to screen; 0 = do not print fragmentation @defragDelay Time to wait between defrag commands; gives the server a little time to catch up @debugMode 1 = display debug comments; helps with troubleshooting 0 = do not display debug comments @rebuildStats Affects only statistics that need to be rebuilt 1 = rebuild stats 0 = do not rebuild stats Called by: SQL Agent Job or DBA Date Initials Version Description ---------------------------------------------------------------------------- 2007-12-18 MFU 1.0 Initial Release 2008-10-17 MFU 1.1 Added @defragDelay, CIX_temp_indexDefragList 2008-11-17 MFU 1.2 Added page_count to log table , added @printFragmentation option 2009-03-17 MFU 2.0 Provided support for centralized execution , consolidated Enterprise & Standard versions , added @debugMode, @maxDopRestriction , modified LOB and partition logic 2009-06-18 MFU 3.0 Fixed bug in LOB logic, added @scanMode option , added support for stat rebuilds (@rebuildStats) , support model and msdb defrag , added columns to the dba_indexDefragLog table , modified logging to show "in progress" defrags , added defrag exclusion list (scheduling) ********************************************************************************* Exec dbo.dba_indexDefrag_sp @executeSQL = 0 , @printCommands = 1 , @debugMode = 1 , @printFragmentation = 1; *********************************************************************************/ SET NOCOUNT ON; SET XACT_Abort ON; SET Ansi_Padding ON; SET Ansi_Warnings ON; SET ArithAbort ON; SET Concat_Null_Yields_Null ON; SET Numeric_RoundAbort OFF; SET Quoted_Identifier ON; BEGIN IF @debugMode = 1 RAISERROR('Undusting the cogs and starting up...', 0, 42) WITH NoWait; /* Declare our variables */ DECLARE @objectID INT , @databaseID INT , @databaseName NVARCHAR(128) , @indexID INT , @partitionCount BIGINT , @schemaName NVARCHAR(128) , @objectName NVARCHAR(128) , @indexName NVARCHAR(128) , @partitionNumber SMALLINT , @fragmentation FLOAT , @pageCount INT , @sqlCommand NVARCHAR(4000) , @rebuildCommand NVARCHAR(200) , @dateTimeStart DATETIME , @dateTimeEnd DATETIME , @containsLOB BIT , @editionCheck BIT , @debugMessage VARCHAR(128) , @updateSQL NVARCHAR(4000) , @partitionSQL NVARCHAR(4000) , @partitionSQL_Param NVARCHAR(1000) , @LOB_SQL NVARCHAR(4000) , @LOB_SQL_Param NVARCHAR(1000) , @rebuildStatsID INT , @rebuildStatsSQL NVARCHAR(1000) , @indexDefrag_id INT; /* Create our temporary tables */ CREATE TABLE #indexDefragList ( databaseID INT , databaseName NVARCHAR(128) , objectID INT , indexID INT , partitionNumber SMALLINT , fragmentation FLOAT , page_count INT , defragStatus BIT , schemaName NVARCHAR(128) Null , objectName NVARCHAR(128) Null , indexName NVARCHAR(128) Null ); CREATE TABLE #databaseList ( databaseID INT , databaseName VARCHAR(128) , scanStatus BIT , statsStatus BIT ); CREATE TABLE #processor ( [INDEX] INT , Name VARCHAR(128) , Internal_Value INT , Character_Value INT ); IF @debugMode = 1 RAISERROR('Beginning validation...', 0, 42) WITH NoWait; /* Just a little validation... */ IF @minFragmentation Not Between 0.00 And 100.0 SET @minFragmentation = 10.0; IF @rebuildThreshold Not Between 0.00 And 100.0 SET @rebuildThreshold = 30.0; IF @defragDelay Not Like '00:[0-5][0-9]:[0-5][0-9]' SET @defragDelay = '00:00:05'; IF @scanMode Not In ('LIMITED', 'SAMPLED', 'DETAILED') SET @scanMode = 'LIMITED'; /* Make sure we're not exceeding the number of processors we have available */ INSERT INTO #processor EXECUTE XP_MSVER 'ProcessorCount'; IF @maxDopRestriction IS Not Null And @maxDopRestriction > (SELECT Internal_Value FROM #processor) SELECT @maxDopRestriction = Internal_Value FROM #processor; /* Check our server version; 1804890536 = Enterprise, 610778273 = Enterprise Evaluation, -2117995310 = Developer */ IF (SELECT SERVERPROPERTY('EditionID')) In (1804890536, 610778273, -2117995310) SET @editionCheck = 1 -- supports online rebuilds ELSE SET @editionCheck = 0; -- does not support online rebuilds IF @debugMode = 1 RAISERROR('Grabbing a list of our databases...', 0, 42) WITH NoWait; /* Retrieve the list of databases to investigate */ INSERT INTO #databaseList SELECT database_id , name , 0 -- not scanned yet for fragmentation , 0 -- statistics not yet updated FROM sys.databases WHERE name = IsNull(@DATABASE, name) And [name] Not In ('master', 'tempdb')-- exclude system databases And [STATE] = 0; -- state must be ONLINE IF @debugMode = 1 RAISERROR('Looping through our list of databases and checking for fragmentation...', 0, 42) WITH NoWait; /* Loop through our list of databases */ WHILE (SELECT COUNT(*) FROM #databaseList WHERE scanStatus = 0) > 0 BEGIN SELECT TOP 1 @databaseID = databaseID FROM #databaseList WHERE scanStatus = 0; SELECT @debugMessage = ' working on ' + DB_NAME(@databaseID) + '...'; IF @debugMode = 1 RAISERROR(@debugMessage, 0, 42) WITH NoWait; /* Determine which indexes to defrag using our user-defined parameters */ INSERT INTO #indexDefragList SELECT database_id AS databaseID , QUOTENAME(DB_NAME(database_id)) AS 'databaseName' , [OBJECT_ID] AS objectID , index_id AS indexID , partition_number AS partitionNumber , avg_fragmentation_in_percent AS fragmentation , page_count , 0 AS 'defragStatus' /* 0 = unprocessed, 1 = processed */ , Null AS 'schemaName' , Null AS 'objectName' , Null AS 'indexName' FROM sys.dm_db_index_physical_stats (@databaseID, OBJECT_ID(@tableName), Null , Null, @scanMode) WHERE avg_fragmentation_in_percent >= @minFragmentation And index_id > 0 -- ignore heaps And page_count > 8 -- ignore objects with less than 1 extent And index_level = 0 -- leaf-level nodes only, supports @scanMode OPTION (MaxDop 2); /* Keep track of which databases have already been scanned */ UPDATE #databaseList SET scanStatus = 1 WHERE databaseID = @databaseID; END CREATE CLUSTERED INDEX CIX_temp_indexDefragList ON #indexDefragList(databaseID, objectID, indexID, partitionNumber); /* Delete any indexes from our to-do that are also in our exclusion list for today */ DELETE idl FROM #indexDefragList AS idl Join dbo.dba_indexDefragExclusion AS ide ON idl.databaseID = ide.databaseID And idl.objectID = ide.objectID And idl.indexID = ide.indexID WHERE exclusionMask & POWER(2, DATEPART(weekday, GETDATE())-1) > 0; SELECT @debugMessage = 'Looping through our list... there''s ' + CAST(COUNT(*) AS VARCHAR(10)) + ' indexes to defrag!' FROM #indexDefragList; IF @debugMode = 1 RAISERROR(@debugMessage, 0, 42) WITH NoWait; /* Begin our loop for defragging */ WHILE (SELECT COUNT(*) FROM #indexDefragList WHERE defragStatus = 0) > 0 BEGIN IF @debugMode = 1 RAISERROR(' Picking an index to beat into shape...', 0, 42) WITH NoWait; /* Grab the most fragmented index first to defrag */ SELECT TOP 1 @objectID = objectID , @indexID = indexID , @databaseID = databaseID , @databaseName = databaseName , @fragmentation = fragmentation , @partitionNumber = partitionNumber , @pageCount = page_count FROM #indexDefragList WHERE defragStatus = 0 ORDER BY fragmentation DESC; IF @debugMode = 1 RAISERROR(' Looking up the specifics for our index...', 0, 42) WITH NoWait; /* Look up index information */ SELECT @updateSQL = N'Update idl Set schemaName = QuoteName(s.name) , objectName = QuoteName(o.name) , indexName = QuoteName(i.name) From #indexDefragList As idl Inner Join ' + @databaseName + '.sys.objects As o On idl.objectID = o.object_id Inner Join ' + @databaseName + '.sys.indexes As i On o.object_id = i.object_id Inner Join ' + @databaseName + '.sys.schemas As s On o.schema_id = s.schema_id Where o.object_id = ' + CAST(@objectID AS VARCHAR(10)) + ' And i.index_id = ' + CAST(@indexID AS VARCHAR(10)) + ' And i.type > 0 And idl.databaseID = ' + CAST(@databaseID AS VARCHAR(10)); EXECUTE SP_EXECUTESQL @updateSQL; /* Grab our object names */ SELECT @objectName = objectName , @schemaName = schemaName , @indexName = indexName FROM #indexDefragList WHERE objectID = @objectID And indexID = @indexID And databaseID = @databaseID; IF @debugMode = 1 RAISERROR(' Grabbing the partition count...', 0, 42) WITH NoWait; /* Determine if the index is partitioned */ SELECT @partitionSQL = 'Select @partitionCount_OUT = Count(*) From ' + @databaseName + '.sys.partitions Where object_id = ' + CAST(@objectID AS VARCHAR(10)) + ' And index_id = ' + CAST(@indexID AS VARCHAR(10)) + ';' , @partitionSQL_Param = '@partitionCount_OUT int OutPut'; EXECUTE SP_EXECUTESQL @partitionSQL, @partitionSQL_Param, @partitionCount_OUT = @partitionCount OUTPUT; IF @debugMode = 1 RAISERROR(' Seeing if there''s any LOBs to be handled...', 0, 42) WITH NoWait; /* Determine if the table contains LOBs */ SELECT @LOB_SQL = ' Select @containsLOB_OUT = Count(*) From ' + @databaseName + '.sys.columns With (NoLock) Where [object_id] = ' + CAST(@objectID AS VARCHAR(10)) + ' And (system_type_id In (34, 35, 99) Or max_length = -1);' /* system_type_id --> 34 = image, 35 = text, 99 = ntext max_length = -1 --> varbinary(max), varchar(max), nvarchar(max), xml */ , @LOB_SQL_Param = '@containsLOB_OUT int OutPut'; EXECUTE SP_EXECUTESQL @LOB_SQL, @LOB_SQL_Param, @containsLOB_OUT = @containsLOB OUTPUT; IF @debugMode = 1 RAISERROR(' Building our SQL statements...', 0, 42) WITH NoWait; /* If there's not a lot of fragmentation, or if we have a LOB, we should reorganize */ IF @fragmentation < @rebuildThreshold Or @containsLOB >= 1 Or @partitionCount > 1 BEGIN SET @sqlCommand = N'Alter Index ' + @indexName + N' On ' + @databaseName + N'.' + @schemaName + N'.' + @objectName + N' ReOrganize'; /* If our index is partitioned, we should always reorganize */ IF @partitionCount > 1 SET @sqlCommand = @sqlCommand + N' Partition = ' + CAST(@partitionNumber AS NVARCHAR(10)); END; /* If the index is heavily fragmented and doesn't contain any partitions or LOB's, rebuild it */ IF @fragmentation >= @rebuildThreshold And IsNull(@containsLOB, 0) != 1 And @partitionCount <= 1 BEGIN /* Set online rebuild options; requires Enterprise Edition */ IF @onlineRebuild = 1 And @editionCheck = 1 SET @rebuildCommand = N' Rebuild With (Online = On'; ELSE SET @rebuildCommand = N' Rebuild With (Online = Off'; /* Set processor restriction options; requires Enterprise Edition */ IF @maxDopRestriction IS Not Null And @editionCheck = 1 SET @rebuildCommand = @rebuildCommand + N', MaxDop = ' + CAST(@maxDopRestriction AS VARCHAR(2)) + N')'; ELSE SET @rebuildCommand = @rebuildCommand + N')'; SET @sqlCommand = N'Alter Index ' + @indexName + N' On ' + @databaseName + N'.' + @schemaName + N'.' + @objectName + @rebuildCommand; END; /* Are we executing the SQL? If so, do it */ IF @executeSQL = 1 BEGIN IF @debugMode = 1 RAISERROR(' Executing SQL statements...', 0, 42) WITH NoWait; /* Grab the time for logging purposes */ SET @dateTimeStart = GETDATE(); /* Log our actions */ INSERT INTO dbo.dba_indexDefragLog ( databaseID , databaseName , objectID , objectName , indexID , indexName , partitionNumber , fragmentation , page_count , dateTimeStart ) SELECT @databaseID , @databaseName , @objectID , @objectName , @indexID , @indexName , @partitionNumber , @fragmentation , @pageCount , @dateTimeStart; SET @indexDefrag_id = SCOPE_IDENTITY(); /* Execute our defrag! */ EXECUTE SP_EXECUTESQL @sqlCommand; SET @dateTimeEnd = GETDATE(); /* Update our log with our completion time */ UPDATE dbo.dba_indexDefragLog SET dateTimeEnd = @dateTimeEnd , durationSeconds = DATEDIFF(SECOND, @dateTimeStart, @dateTimeEnd) WHERE indexDefrag_id = @indexDefrag_id; /* Just a little breather for the server */ WAITFOR Delay @defragDelay; /* Print if specified to do so */ IF @printCommands = 1 PRINT N'Executed: ' + @sqlCommand; END ELSE /* Looks like we're not executing, just printing the commands */ BEGIN IF @debugMode = 1 RAISERROR(' Printing SQL statements...', 0, 42) WITH NoWait; IF @printCommands = 1 PRINT IsNull(@sqlCommand, 'error!'); END IF @debugMode = 1 RAISERROR(' Updating our index defrag status...', 0, 42) WITH NoWait; /* Update our index defrag list so we know we've finished with that index */ UPDATE #indexDefragList SET defragStatus = 1 WHERE databaseID = @databaseID And objectID = @objectID And indexID = @indexID And partitionNumber = @partitionNumber; END /* Do we want to output our fragmentation results? */ IF @printFragmentation = 1 BEGIN IF @debugMode = 1 RAISERROR(' Displaying fragmentation results...', 0, 42) WITH NoWait; SELECT databaseID , databaseName , objectID , objectName , indexID , indexName , fragmentation , page_count FROM #indexDefragList; END; /* Do we want to rebuild stats? */ IF @rebuildStats = 1 BEGIN WHILE Exists(SELECT TOP 1 * FROM #databaseList WHERE statsStatus = 0) BEGIN /* Build our SQL statement to update stats */ SELECT TOP 1 @rebuildStatsSQL = 'Use [' + databaseName + ']; ' + 'Execute sp_updatestats;' , @rebuildStatsID = databaseID FROM #databaseList WHERE statsStatus = 0; SET @debugMessage = 'Rebuilding Statistics: ' + @rebuildStatsSQL; IF @debugMode = 1 RAISERROR(@debugMessage, 0, 42) WITH NoWait; /* Execute our stats update! */ EXECUTE SP_EXECUTESQL @rebuildStatsSQL; /* Keep track of which databases have been updated */ UPDATE #databaseList SET statsStatus = 1 WHERE databaseID = @rebuildStatsID; END; END; /* When everything is said and done, make sure to get rid of our temp table */ DROP TABLE #indexDefragList; DROP TABLE #databaseList; DROP TABLE #processor; IF @debugMode = 1 RAISERROR('DONE! Thank you for taking care of your indexes! :)', 0, 42) WITH NoWait; SET NOCOUNT OFF; RETURN 0 END Go SET Quoted_Identifier OFF SET ANSI_Nulls ON Go
Overhead in Non-Unique Clustered Indexes
Filed under: Internals, Performance & Tuning, SQL 2008, Syndication
I’ve received a couple of questions regarding my article, Performance Considerations of Data Types, and the overhead associated with non-unique clustered indexes. I started to respond via e-mail, but my response was so long I decided to turn it into a blog post instead. 
I should start by clarifying that non-unique clustered indexes do not necessarily consume more space and overhead; it depends on the data stored. If you have duplicate clustered key values, the first instance of the value will be handled as though it were unique. Any subsequent values, however, will incur overhead to manage the uniquifier that SQL Server adds to maintain row uniqueness. This same overhead is also incurred in non-clustered indexes, too, adding to the overall expense of this approach.
I think it helps to actually look at the data, so let’s walk through a few different common scenarios. We’ll create a table with a unique clustered index, a table with a non-unique clustered index but no duplicates, and a table with duplicate key values.
Also, a little warning that I started to write this in SQL Server 2008, and since I’m on a 2008 kick, I decided to leave it that way. You can modify this pretty easily to work in 2005, if necessary.
USE sandbox; Go /* Unique, clustered index, no duplicate values */ CREATE TABLE dbo.uniqueClustered ( myDate DATE Not Null , myNumber INT Not Null , myColumn CHAR(995) Not Null ); CREATE UNIQUE CLUSTERED INDEX CIX_uniqueClustered ON dbo.uniqueClustered(myDate); /* Non-unique clustered index, but no duplicate values */ CREATE TABLE dbo.nonUniqueNoDups ( myDate DATE Not Null , myNumber INT Not Null , myColumn CHAR(995) Not Null ); CREATE CLUSTERED INDEX CIX_nonUniqueNoDups ON dbo.nonUniqueNoDups(myDate); /* Non-unique clustered index, duplicate values */ CREATE TABLE dbo.nonUniqueDuplicates ( myDate DATE Not Null , myNumber INT Not Null , myColumn CHAR(995) Not Null ); CREATE CLUSTERED INDEX CIX_nonUniqueDuplicates ON dbo.nonUniqueDuplicates(myDate);
I’m going to use the date data type in 2008 for my clustered index key. To ensure uniqueness for the first two tables, I’ll iterate through a few years’ worth of dates. This is typical of what you may see in a data mart, where you’d have one record with an aggregation of each day’s data. For the table with duplicate values, I’m going to insert the same date for each row.
/* Populate some test data */ SET NOCOUNT ON; DECLARE @myDate DATE = '1990-01-01' , @myNumber INT = 1; WHILE @myDate < '2010-01-01' BEGIN INSERT INTO dbo.uniqueClustered SELECT @myDate, @myNumber, 'data'; INSERT INTO dbo.nonUniqueNoDups SELECT @myDate, @myNumber, 'data'; INSERT INTO dbo.nonUniqueDuplicates SELECT '2009-01-01', @myNumber, 'data'; SELECT @myDate = DATEADD(DAY, 1, @myDate) , @myNumber += 1; END;
After running the above script, each table should have 7,305 records. This is obviously pretty small for a table, but it’ll serve our purposes. Now let’s take a look at the size of our tables:
/* Look at the details of our indexes */ /* Unique, clustered index, no duplicate values */ SELECT 'unique' AS 'type', page_count, avg_page_space_used_in_percent, record_count , min_record_size_in_bytes, max_record_size_in_bytes FROM sys.dm_db_index_physical_stats(DB_ID(), OBJECT_ID(N'uniqueClustered'), Null, Null, N'Detailed') WHERE index_level = 0 UNION All /* Non-unique clustered index, but no duplicate values */ SELECT 'non-unique, no dups', page_count, avg_page_space_used_in_percent, record_count , min_record_size_in_bytes, max_record_size_in_bytes FROM sys.dm_db_index_physical_stats(DB_ID(), OBJECT_ID(N'nonUniqueNoDups'), Null, Null, N'Detailed') WHERE index_level = 0 UNION All /* Non-unique clustered index, duplicate values */ SELECT 'duplicates', page_count, avg_page_space_used_in_percent, record_count , min_record_size_in_bytes, max_record_size_in_bytes FROM sys.dm_db_index_physical_stats(DB_ID(), OBJECT_ID(N'nonUniqueDuplicates'), Null, Null, N'Detailed') WHERE index_level = 0;
Here’s the results:
type page_count avg_page_space_used_in_percent record_count min_record_size_in_bytes max_record_size_in_bytes ------------------- -------------------- ------------------------------ -------------------- ------------------------ ------------------------ unique 914 99.8055102545095 7305 1009 1009 non-unique, no dups 914 99.8055102545095 7305 1009 1009 duplicates 1044 88.066036570299 7305 1009 1017
I want to point out a couple of things. First, there is no difference in the number of pages between the non-unique clustered index with no duplicates ([nonUniqueNoDups]) and the unique clustered index ([uniqueClustered]). The table with duplicate clustered key values, however, requires 14% more pages to store the same amount of data. Secondly, the [max_record_size_in_bytes] of the [nonUniqueDuplicates] table is 8 bytes more than that of the other two. We’ll discuss why in a minute.
Now let’s take a look at the actual data pages. For this, I’m going to use my page internals proc.
Execute dbo.dba_viewPageData_sp @databaseName = 'sandbox' , @tableName = 'sandbox.dbo.uniqueClustered' , @indexName = 'CIX_uniqueClustered';
I’m not going to post the entire results here, but I want to draw your attention to “m_slotCnt = 8″, which is near the top of the page. That means 8 records are stored on this page. Also, when you look near the end of the first record (Slot 0), you should see the following results:
Slot 0 Offset 0x60 Length 1009 Record Type = PRIMARY_RECORD Record Attributes = NULL_BITMAP Record Size = 1009 Memory Dump @0x00A9C060 00000000: 1000ee03 c3150b01 00000064 61746120 †..î.Ã......data [...] 000003F0: 00†††††††††††††††††††††††††††††††††††. Slot 0 Column 1 Offset 0x4 Length 3 Length (physical) 3 myDate = 1990-01-01 Slot 0 Column 2 Offset 0x7 Length 4 Length (physical) 4 myNumber = 1 Slot 0 Column 3 Offset 0xb Length 995 Length (physical) 995 myColumn = data
Now let’s look at the table that has a non-unique clustered index but no duplicates:
EXECUTE dbo.dba_viewPageData_sp @databaseName = 'sandbox' , @tableName = 'sandbox.dbo.nonUniqueNoDups' , @indexName = 'CIX_nonUniqueNoDups';
The m_slotCnt count is also 8 for this page. This time, let’s glance at the first and second records (Slot 0 and 1 respectively):
Slot 0 Offset 0x60 Length 1009 Record Type = PRIMARY_RECORD Record Attributes = NULL_BITMAP Record Size = 1009 Memory Dump @0x62FDC060 00000000: 1000ee03 c3150b01 00000064 61746120 †..î.Ã......data [...] 000003F0: 00†††††††††††††††††††††††††††††††††††. Slot 0 Column 0 Offset 0x0 Length 4 Length (physical) 0 UNIQUIFIER = 0 Slot 0 Column 1 Offset 0x4 Length 3 Length (physical) 3 myDate = 1990-01-01 Slot 0 Column 2 Offset 0x7 Length 4 Length (physical) 4 myNumber = 1 Slot 0 Column 3 Offset 0xb Length 995 Length (physical) 995 myColumn = data Slot 1 Offset 0x451 Length 1009 Record Type = PRIMARY_RECORD Record Attributes = NULL_BITMAP Record Size = 1009 Memory Dump @0x62FDC451 00000000: 1000ee03 c4150b02 00000064 61746120 †..î.Ä......data [...] 000003F0: 00†††††††††††††††††††††††††††††††††††. Slot 1 Column 0 Offset 0x0 Length 4 Length (physical) 0 UNIQUIFIER = 0 Slot 1 Column 1 Offset 0x4 Length 3 Length (physical) 3 myDate = 1990-01-02 Slot 1 Column 2 Offset 0x7 Length 4 Length (physical) 4 myNumber = 2 Slot 1 Column 3 Offset 0xb Length 995 Length (physical) 995 myColumn = data
We now see a new addition to the row, “UNIQUIFIER = 0.” This is SQL Server’s way of managing row uniqueness internally. You’ll notice that, because the clustered key values are unique, the UNIQUIFIER is set to 0 and the row size is still 1009; for all intents and purposes, the UNIQUIFIER is not consuming any space.
Update: The DBCC God himself, Paul Randal, explained that non-dupes actually have a NULL UNIQUIFIER, which DBCC PAGE displays as a 0. Thanks for explaining, Paul! I wondered about that but chalked it up to SQL voodoo.
Now let’s look at our final case, a non-unique clustered index with duplicate key values:
EXECUTE dbo.dba_viewPageData_sp @databaseName = 'sandbox' , @tableName = 'sandbox.dbo.nonUniqueDuplicates' , @indexName = 'CIX_nonUniqueDuplicates';
Here’s where things get interesting. The m_slotCnt value is now 7, which means we’re now storing 1 record less per page. Let’s look at the details:
Slot 0 Offset 0x60 Length 1009 Record Type = PRIMARY_RECORD Record Attributes = NULL_BITMAP Record Size = 1009 Memory Dump @0x00A9C060 00000000: 1000ee03 df300b01 00000064 61746120 †..î.ß0.....data [...] 000003F0: 00†††††††††††††††††††††††††††††††††††. Slot 0 Column 0 Offset 0x0 Length 4 Length (physical) 0 UNIQUIFIER = 0 Slot 0 Column 1 Offset 0x4 Length 3 Length (physical) 3 myDate = 2009-01-01 Slot 0 Column 2 Offset 0x7 Length 4 Length (physical) 4 myNumber = 1 Slot 0 Column 3 Offset 0xb Length 995 Length (physical) 995 myColumn = data Slot 1 Offset 0x451 Length 1017 Record Type = PRIMARY_RECORD Record Attributes = NULL_BITMAP VARIABLE_COLUMNS Record Size = 1017 Memory Dump @0x00A9C451 00000000: 3000ee03 df300b02 00000064 61746120 †0.î.ß0.....data [...] 000003F0: 000100f9 03010000 00†††††††††††††††††...ù..... Slot 1 Column 0 Offset 0x3f5 Length 4 Length (physical) 4 UNIQUIFIER = 1 Slot 1 Column 1 Offset 0x4 Length 3 Length (physical) 3 myDate = 2009-01-01 Slot 1 Column 2 Offset 0x7 Length 4 Length (physical) 4 myNumber = 2 Slot 1 Column 3 Offset 0xb Length 995 Length (physical) 995 myColumn = data
The first record, Slot 0, looks exactly the same as in the previous table; the UNIQUIFIER is 0 and the row size is 1009. The second record (Slot 1), however, now has a UNIQUIFIER value of 1 and the row size is 1017. If you notice, the “Record Attributes” of Slot 1 are also different, with the addition of “VARIABLE_COLUMNS.” This is because the UNIQUIFIER is stored as a variable column. The extra 8 bytes of overhead break down to 4 bytes to store the UNIQUIFIER, 2 bytes to store the variable column offset, and 2 bytes to store the variable count. The tables we created used all fixed-length columns; you may notice some minor overhead differences if your table already contains variable columns.
To summarize, there is indeed a difference in the page structure between a unique clustered index and a non-unique clustered index; however, there’s only a possible performance and space impact when storing duplicate clustered key values. So there you go, more detail than you ever wanted to know about clustered indexes and uniqueness!
Performance Considerations of Data Types
Filed under: PASS, Performance & Tuning, SQL Tips, Syndication
I’ve just finished my first real content for the PASS Performance SIG. I decided to write on “Performance Considerations of Data Types,” as I think this is one of the easiest and most overlooked topics in performance tuning. Here’s a summary:
Selecting inappropriate data types, especially on large tables with millions or billions of rows, can have significant performance implications. In this article, I’ll explain why and offer suggestions on how to select the most appropriate data type for your needs. The primary focus will be on common data types in SQL Server 2005 and 2008, but I’ll also discuss some aspects of clustered indexes and column properties. Most importantly, I’ll show some examples of common data-type misuse.
If you’re interested in this content, you can find it here: Performance Considerations of Data Types.
Special thanks to Paul Randal and Paul Nielsen for providing me with technical reviews and great feedback. You guys are awesome!
Thanks also to Mladen Prajdic and Jeremiah Peschka for their great input. You guys are awesome, too!
A Look at Missing Indexes
Filed under: Performance & Tuning, Syndication, T-SQL Scripts
Tim Ford (@SQLAgentMan) recently blogged about his Top 5 SQL Server Indexing Best Practices. I thought it was a good list, and it inspired this blog post. I’ve recently been doing a little index spring cleaning, and I thought some people may be interested in the process I go through. So, here it is… a journey through madness an overview of my general missing index process.
I start with my trusty dba_missingIndexStoredProc table. If this table sounds completely foreign to you, check out my post, Find Missing Indexes In Stored Procs. Basically, I have a process that runs every night, scanning the XML of every query plan on the server to find procs that are possibly missing indexes. I then log the details for later action.
So I take a look at my table, and I find 8 stored procedures that are possibly missing indexes. Clicking on the XML link will show me the logged query plan:
Procs With Missing Indexes
Right clicking on the “Missing Index” description will give me the details of the recommended index:
Missing Index Details
Here’s an example of what SQL Server will return for you:
/* Missing Index Details from ExecutionPlan2.sqlplan The Query Processor estimates that implementing the following index could improve the query cost by 85.7327%. */ /* USE [msdb] GO CREATE NONCLUSTERED INDEX [<Name of Missing Index, sysname,>] ON [dbo].[sysjobhistory] ([job_id]) INCLUDE ([instance_id],[step_id],[sql_message_id],[sql_severity], [run_status],[run_date],[run_time],[run_duration],[operator_id_emailed], [operator_id_netsent],[operator_id_paged],[retries_attempted],[server]) GO */
I now compare the details of this proposed index to the missing index DMV suggestions, using this query:
SELECT t.name AS 'affected_table' , 'Create NonClustered Index IX_' + t.name + '_missing_' + CAST(ddmid.index_handle AS VARCHAR(10)) + ' On ' + ddmid.STATEMENT + ' (' + IsNull(ddmid.equality_columns,'') + CASE WHEN ddmid.equality_columns IS Not Null And ddmid.inequality_columns IS Not Null THEN ',' ELSE '' END + IsNull(ddmid.inequality_columns, '') + ')' + IsNull(' Include (' + ddmid.included_columns + ');', ';' ) AS sql_statement , ddmigs.user_seeks , ddmigs.user_scans , CAST((ddmigs.user_seeks + ddmigs.user_scans) * ddmigs.avg_user_impact AS INT) AS 'est_impact' , ddmigs.last_user_seek FROM sys.dm_db_missing_index_groups AS ddmig INNER JOIN sys.dm_db_missing_index_group_stats AS ddmigs ON ddmigs.group_handle = ddmig.index_group_handle INNER JOIN sys.dm_db_missing_index_details AS ddmid ON ddmig.index_handle = ddmid.index_handle INNER JOIN sys.tables AS t ON ddmid.OBJECT_ID = t.OBJECT_ID WHERE ddmid.database_id = DB_ID() --AND t.name = 'myTableName' ORDER BY CAST((ddmigs.user_seeks + ddmigs.user_scans) * ddmigs.avg_user_impact AS INT) DESC;
I usually find the data in both places, but not always. One reason why is because the missing index DMV will only store data since your last reboot. So if I’m taking a look at this DMV on Monday and I just rebooted on Sunday, I may not have enough history to give me meaningful recommendations. This is just something to be aware of.
What I’m looking for in this DMV is the number of user_seeks and the est_impact. Also, if I haven’t rebooted my server in a while, I take a look at last_user_seek so I can determine whether or not it’s still accurate.
Next, I take a look at my existing indexes using Kimberly Tripp’s sp_helpindex2 system stored proc. I use her proc instead of sp_helpindex because I need to see included columns.
If you’re wondering why I’m looking at existing indexes, the reason is because I’m looking for indexes that can be modified slightly to accommodate my missing index needs. By “modified slightly,” I mean that I’d only want to make a change to an existing index if it did not drastically change the size or composition of an index, i.e. adding one or two narrow columns as included columns. I do NOT mean making changes that double the size of your index; in those cases, you’d probably be better off creating a brand new index.
Looking at existing indexes is actually a pretty critical part of the puzzle. If I have a proc that only gets called a few times an hour and could benefit from a better index, I may not create that index if it means adding a wide, expensive index to a busy table. If I can make a small modification to an existing index, then there’s a greater chance I’ll make the change and cover my query.
At this point, I should have enough information to start making decisions. I was going to write out the path I normally take when making decisions, but I thought, “Hey! What a great time for a diagram.” So here you go:
Decision Path
Disclaimer: I’m *not* a Visio wizard, so if I butchered the use of certain symbols in my diagram, please let me know so I can a) fix it, and b) learn from it!
It’s hard to really put all of the decision paths into a single, small diagram like this. There’s a lot of variables that I’m not even touching here. But I think this is a fairly good “generic” representation of the path I take. When I hit an “end” process, it means I don’t create the missing index at this time. Maybe in the future, it’ll become necessary, but I prefer to err on the side of less indexes.
So there you have it, a brief look at my missing index process. Hopefully someone finds it helpful.
Indexing for Partitioned Tables
So you’ve partitioned your table, and now you’re ready to performance tune. As with any table, indexing is a great place to start. And if you’re like most people new to partitioning, you probably created all of your indexes on the partitioned scheme, either by design or unintentionally.
Let’s take a step back here and discuss what a partitioned index is. A partitioned index, like a partitioned table, separates data into different physical structures (partitions) under one logical name. Specifically, each partition in a nonclustered index contains its own B-tree structure with a subset of rows, based upon the partitioning scheme. By default, an unpartitioned nonclustered index has just one partition.
Keep in mind, when you create an index on a partitioned table, i.e.
CREATE NONCLUSTERED INDEX IX_myIndex ON dbo.myTable(myColumn);
… you are creating the index on the partitioned scheme by default. In order to create a NON-partitioned index on that same table, you would need to explicitly declare “On [FileGroup]“, i.e.
CREATE NONCLUSTERED INDEX IX_myIndex ON dbo.myTable(myColumn) ON [PRIMARY];
But should you partition your index? That depends on how you use it. In fact, most environments will probably want to use a mix of partitioned and non-partitioned indexes. I’ve found that that partitioned indexes perform better when aggregating data or scanning partitions. Conversely, you’ll probably find that, if you need to locate a single, specific record, nothing performs better than a non-partitioned index on that column.
Let’s walk through some examples and see how they perform. I’ll bring back my trusty ol’ orders table for this.
/* Create a partition function. */ CREATE Partition FUNCTION [test_monthlyDateRange_pf] (DATETIME) AS Range RIGHT FOR VALUES ('2009-01-01', '2009-01-08', '2009-01-15' , '2009-01-22', '2009-01-29'); Go /* Associate the partition function with a partition scheme. */ CREATE Partition Scheme test_monthlyDateRange_ps AS Partition test_monthlyDateRange_pf All TO ([PRIMARY]); Go /* Create a partitioned table. */ CREATE TABLE dbo.orders ( order_id INT IDENTITY(1,1) Not Null , orderDate DATETIME Not Null , orderData SMALLDATETIME Not Null CONSTRAINT PK_orders PRIMARY KEY CLUSTERED ( order_id , orderDate ) ) ON test_monthlyDateRange_ps(orderDate); Go /* Create some records to play with. */ SET NOCOUNT ON; DECLARE @endDate DATETIME = '2009-01-01'; WHILE @endDate < '2009-02-01' BEGIN INSERT INTO dbo.orders SELECT @endDate, @endDate; SET @endDate = DATEADD(MINUTE, 1, @endDate); END; SET NOCOUNT OFF; /* Let’s create an aligned, partitioned index. */ CREATE NONCLUSTERED INDEX IX_orders_aligned ON dbo.orders(order_id) ON test_monthlyDateRange_ps(orderDate); /* you don't actually need to declare the last line of this unless you want to create the index on a different partitioning scheme. */ /* Now let’s create an unpartitioned index. */ CREATE NONCLUSTERED INDEX IX_orders_unpartitioned ON dbo.orders(order_id) ON [PRIMARY];
Now that we have both a partitioned and an unpartitioned index, let’s take a look at our sys.partitions table:
/* Let's take a look at our index partitions */ SELECT i.name , i.index_id , p.partition_number , p.ROWS FROM sys.partitions AS p Join sys.indexes AS i ON p.OBJECT_ID = i.OBJECT_ID And p.index_id = i.index_id WHERE p.OBJECT_ID = OBJECT_ID('orders') ORDER BY i.index_id, p.partition_number;
sys.partitions
As expected, both of our partitioned indexes, PK_orders and IX_orders_aligned, have 6 partitions, with a subset of rows on each partition. Our unpartitioned non-clustered index, IX_orders_unpartitioned, on the other hand has just 1 partition containing all of the rows.
Now that we have our environment set up, let’s run through some different queries and see the performance impact of each type of index.
/* Query 1, specific record look-up, covered */ SELECT order_id, orderDate FROM dbo.orders WITH (INDEX(IX_orders_aligned)) WHERE order_id = 25000; SELECT order_id, orderDate FROM dbo.orders WITH (INDEX(IX_orders_unpartitioned)) WHERE order_id = 25000;
Query 1
The unpartitioned index performs significantly better when given a specific record to look-up. Now let’s try the same query, but utilizing a scan instead of a seek:
/* Query 2, specific record look-up, uncovered */ SELECT order_id, orderDate, orderData FROM dbo.orders WITH (INDEX(IX_orders_aligned)) WHERE order_id = 30000; SELECT order_id, orderDate, orderData FROM dbo.orders WITH (INDEX(IX_orders_unpartitioned)) WHERE order_id = 30000;
Query 2
Again we see that the non-partitioned index performs better with the single-record look-up. This can lead to some pretty dramatic performance implications. So when *would* we want to use a partitioned index? Two instances immediately pop to mind. First, partition switching can only be performed when all indexes on a table are aligned. Secondly, partitioned indexes perform better when manipulating large data sets. To see this in action, let’s try some simple aggregation…
/* Query 3, aggregation */ SELECT CAST(ROUND(CAST(orderdate AS FLOAT), 0, 1) AS SMALLDATETIME) AS 'order_date' , COUNT(*) FROM dbo.orders WITH (INDEX(IX_orders_aligned)) WHERE orderDate Between '2009-01-01' And '2009-01-07 23:59' GROUP BY CAST(ROUND(CAST(orderdate AS FLOAT), 0, 1) AS SMALLDATETIME) ORDER BY CAST(ROUND(CAST(orderdate AS FLOAT), 0, 1) AS SMALLDATETIME); SELECT CAST(ROUND(CAST(orderdate AS FLOAT), 0, 1) AS SMALLDATETIME) AS 'order_date' , COUNT(*) FROM dbo.orders WITH (INDEX(IX_orders_unpartitioned)) WHERE orderDate Between '2009-01-01' And '2009-01-07 23:59' GROUP BY CAST(ROUND(CAST(orderdate AS FLOAT), 0, 1) AS SMALLDATETIME) ORDER BY CAST(ROUND(CAST(orderdate AS FLOAT), 0, 1) AS SMALLDATETIME);
Query 3
As you can see, partitioned indexes perform better when aggregating data. This is just a simple example, but the results can be even more dramatic in a large production environment. This is one of the reasons why partitioned tables and indexes are especially beneficial in data warehouses.
So now you have a general idea of what a partitioned index is and when to use a partitioned index vs a non-partitioned index. Ultimately, your indexing needs will depend largely on the application and how the data is used. When in doubt, test, test, test! So to recap…
- Specify “On [FileGroup]“ to create an unpartitioned index on a partitioned table
- Consider using non-partitioned indexes for single-record look-ups
- Use partitioned indexes for multiple records and data aggregations
- To enable partition switching, all indexes on the table must be aligned.
For more information on partitioning, check out my other partitioning articles:
Partitioning Example
Partitioning 101
Tips for Large Data Stores
Subscribe to my blog!
It's easy, free, and will keep you informed of my latest SQL tips and scripts.Like what you see? Subscribe! 
Around the Web
SQLRockstar Blog Ranking
