Sql-server – SQL Server nvarchar(max) vs nvarchar(n) affects performance

sql serversql-server-2008-r2

This is SQL Server 2008 R2 SP2. I have 2 tables. Both are identical (data and indexing), except the first table has a VALUE column as nvarchar(max) and the second has the same column as nvarchar(800). This column is included in a non-clustered index. I also created a clustered index on both tables. I have also rebuilt the indexes. The max string length in this column is 650.

If I run the same query against both the nvarchar(800) table is consistently faster, many times twice as fast. Sure seems like it is defeating the purpose of "varchar". Table contains 800,000+ rows. The query should be looking at around 110,000 rows (which is what the plan estimates).

According to the io stats there are no lob reads, so everything appears to be in row. Execution plans are the same, except there is a slight difference in the cost percentage between the two tables and the estimated row size is bigger with the nvarchar(max) (91 bytes vs 63 bytes). The number of reads are pretty much the same too.

Why the difference?

===== Schema ======

 CREATE TABLE [dbo].[table1](
        [ID] [bigint] IDENTITY(1,1) NOT NULL,
        [ProductID] [bigint] NOT NULL,
        [ProductSkeletonID] [bigint] NOT NULL,
        [Value] [nvarchar](max) NOT NULL,
        [IsKeywordSearchable] [bit] NULL,
        [ValueInteger] [bigint] NULL,
        [ValueDecimal] [decimal](18, 2) NULL,
        [ValueDate] [datetime] NULL,
        [TypeOfData] [nvarchar](20) NOT NULL,
     CONSTRAINT [PK_table1] PRIMARY KEY CLUSTERED 
    (
        [ID] ASC
    )WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]
    ) ON [PRIMARY] TEXTIMAGE_ON [PRIMARY]

    CREATE NONCLUSTERED INDEX [IX_table1_productskeletonid] ON [dbo].[table1] 
    (
        [ProductSkeletonID] ASC
    )
    INCLUDE ( [ProductID],
    [Value]) WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]

    CREATE TABLE [dbo].[table2](
        [ID] [bigint] IDENTITY(1,1) NOT NULL,
        [ProductID] [bigint] NOT NULL,
        [ProductSkeletonID] [bigint] NOT NULL,
        [Value] [nvarchar](800) NOT NULL,
        [IsKeywordSearchable] [bit] NULL,
        [ValueInteger] [bigint] NULL,
        [ValueDecimal] [decimal](18, 2) NULL,
        [ValueDate] [datetime] NULL,
        [TypeOfData] [nvarchar](20) NOT NULL,
     CONSTRAINT [PK_table2] PRIMARY KEY CLUSTERED 
    (
        [ID] ASC
    )WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]
    ) ON [PRIMARY]

    CREATE NONCLUSTERED INDEX [IX_table2_productskeletonid] ON [dbo].[table2] 
    (
        [ProductSkeletonID] ASC
    )
    INCLUDE ( [ProductID],
    [Value]) WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]


CREATE TABLE [dbo].[table_results](
    [SearchID] [bigint] NOT NULL,
    [RowNbr] [int] NOT NULL,
    [ProductID] [bigint] NOT NULL,
    [PermissionList] [varchar](250) NULL,
    [SearchWeight] [int] NULL,
 CONSTRAINT [PK_table_results] PRIMARY KEY NONCLUSTERED 
(
    [SearchID] ASC,
    [RowNbr] ASC
)WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]
) ON [PRIMARY]

CREATE NONCLUSTERED INDEX [IX_table_results_SearchID] ON [dbo].[cart_product_searches_results] 
(
    [SearchID] ASC
)
INCLUDE ( [ProductID]) WITH (PAD_INDEX  = OFF, STATISTICS_NORECOMPUTE  = OFF, SORT_IN_TEMPDB = OFF, IGNORE_DUP_KEY = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS  = ON, ALLOW_PAGE_LOCKS  = ON) ON [PRIMARY]

===== Table1 query ======

    SELECT cppev.ProductSkeletonID, cppev.Value, COUNT(*) AS Value FROM table1 cppev
    JOIN search_results cpsr ON cppev.ProductID = cpsr.ProductID AND cpsr.SearchID = 227568 
    WHERE cppev.ProductSkeletonID in (3191, 3160, 3158, 3201)
    GROUP BY cppev.ProductSkeletonID, cppev.Value

    Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
    Table 'table1'. Scan count 4, logical reads 582, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
    Table 'table_results'. Scan count 1, logical reads 82, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

    SQL Server Execution Times:
       CPU time = 1373 ms,  elapsed time = 1576 ms.

 |--Compute Scalar(DEFINE:([Expr1005]=CONVERT_IMPLICIT(int,[Expr1008],0)))
       |--Stream Aggregate(GROUP BY:([cppev].[Value], [cppev].[ProductSkeletonID]) DEFINE:([Expr1008]=Count(*)))
            |--Sort(ORDER BY:([cppev].[Value] ASC, [cppev].[ProductSkeletonID] ASC))
                 |--Hash Match(Inner Join, HASH:([cpsr].[ProductID])=([cppev].[ProductID]), RESIDUAL:([dbo].[table1].[ProductID] as [cppev].[ProductID]=[dbo].[table_results].[ProductID] as [cpsr].[ProductID]))
                      |--Index Seek(OBJECT:([dbo].[table_results].[IX_table_results_SearchID] AS [cpsr]), SEEK:([cpsr].[SearchID]=(227568)) ORDERED FORWARD)
                      |--Index Seek(OBJECT:([dbo].[table1].[IX_table1_productskeletonid] AS [cppev]), SEEK:([cppev].[ProductSkeletonID]=(3158) OR [cppev].[ProductSkeletonID]=(3160) OR [cppev].[ProductSkeletonID]=(3191) OR [cppev].[ProductSkeletonID]=(3201)) ORDERED FORWARD)

===== Table2 query ======

    SELECT cppev.ProductSkeletonID, cppev.Value, COUNT(*) AS Value FROM table2 cppev
    JOIN table_results cpsr ON cppev.ProductID = cpsr.ProductID AND cpsr.SearchID = 227568 
    WHERE cppev.ProductSkeletonID in (3191, 3160, 3158, 3201)
    GROUP BY cppev.ProductSkeletonID, cppev.Value

    Table 'Worktable'. Scan count 0, logical reads 0, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
    Table 'table2'. Scan count 4, logical reads 584, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.
    Table 'table_results'. Scan count 1, logical reads 82, physical reads 0, read-ahead reads 0, lob logical reads 0, lob physical reads 0, lob read-ahead reads 0.

    SQL Server Execution Times:
       CPU time = 484 ms,  elapsed time = 796 ms.

  |--Compute Scalar(DEFINE:([Expr1005]=CONVERT_IMPLICIT(int,[Expr1008],0)))
       |--Stream Aggregate(GROUP BY:([cppev].[Value], [cppev].[ProductSkeletonID]) DEFINE:([Expr1008]=Count(*)))
            |--Sort(ORDER BY:([cppev].[Value] ASC, [cppev].[ProductSkeletonID] ASC))
                 |--Hash Match(Inner Join, HASH:([cpsr].[ProductID])=([cppev].[ProductID]), RESIDUAL:([auctori_core_v40_D].[dbo].[table2].[ProductID] as [cppev].[ProductID]= [dbo].[table2].[ProductID] as [cpsr].[ProductID]))
                      |--Index Seek(OBJECT:([dbo].[table_results].[IX_table_results_SearchID] AS [cpsr]), SEEK:([cpsr].[SearchID]=(227568)) ORDERED FORWARD)
                      |--Index Seek(OBJECT:([dbo].[table2].[IX_table2_productskeletonid] AS [cppev]), SEEK:([cppev].[ProductSkeletonID]=(3158) OR [cppev].[ProductSkeletonID]=(3160) OR [cppev].[ProductSkeletonID]=(3191) OR [cppev].[ProductSkeletonID]=(3201)) ORDERED FORWARD)

Best Answer

You're seeing the cost overhead of using MAX types.

While NVARCHAR(MAX) is identical to NVARCHAR(n) in TSQL and can be stored in-row, it is handled separately by the storage engine because it can be pushed off-row. When off-row it is a LOB_DATA allocation unit, rather than ROW_OVERFLOW_DATA allocation unit and we can assume from your observations that this carries an overhead.

You can see the two types are internally stored differently with a little DBCC PAGE spelunking. Mark Rasmussen posted example page dumps that show the differences in What is the Size of the LOB Pointer for (MAX) Types Like Varchar, Varbinary, Etc?

We can probably assume it's the GROUP BY on the MAX column that causes the performance difference in your case. I've not tested other operations on a MAX type but it might be interesting to do so and see if similar results are seen.

Related Solutions

Sql-server – sql server partitioned view execution plan questions

The DATEPART() function still has to be performed on every row before any filtering can be done/calculated.

Your view would be much more efficient if you changed it like this:

CREATE VIEW PartitionedTb_Sales 
AS 
SELECT *, 2010 AS SalesYear FROM tb_sales2010
UNION ALL
SELECT *, 2011 AS SalesYear FROM tb_sales2011
UNION ALL
SELECT *, 2012 AS SalesYear FROM tb_sales2012

Then, instead of using DATEPART() to calculate the year for each row, it will already be defined.

SELECT TOP 1 *
FROM partitionedTb_sales
WHERE SalesYear = 2011

Sql-server – Which of these queries is best for performance

Sometimes I wonder if SHORT scripts really is the best thing to focus on.

The size of a script has little to do with how efficiently the query will execute. A more compact statement will likely consume fewer resources in terms of compilation, but (re)compilation is usually a rare occurrence in a live system.

Fewer table accesses is usually desirable, though, and this does lead to more compact code.

Very generally speaking, a smaller execution plan will yield better results, and a lower estimated cost will yield better results. Again, though, it's highly situational. Cost estimates in particular can be way off in some cases. It's important to measure the actual execution time, because at the end of the day, that's what matters.

With left joins i can achieve what i want with just a few lines. But then I tried with a longer script, using unions. Which is the best method?

First of all, we need to know how much data will be in these tables in a real system. Right now there's so little it will be difficult to use the STATISTICS TIME performance metrics to figure out a winner -- the results that come back will be dominated by factors other than the query execution. With more data, it's likely the plans will change, thus rendering the comparison here moot.

Having said that, by looking at the query plans as they are now from a logical point of view, the first one is the winner.

You can see that the Clustered Index Scan of quantities appears once in the first plan, while it appears four times in the second one. The second plan also contains an expensive Distinct Sort as a result of using UNIONs (this operator could be eliminated by using UNION ALLs instead, which won't change the results).

The first query could also probably be improved, by getting index seeks on the colors and sizes tables, instead of table scans. It might be worth trying a hash match plan as well (which is what you'll probably see when quantities and products are larger), but for tables this small, the startup cost may be too much overhead to be of benefit.

What I would suggest you do is run each of the statements you want to test 10,000+ times in a loop, figure out the average execution time, and then compare.

Best Answer

Related Solutions

Sql-server – sql server partitioned view execution plan questions

Sql-server – Which of these queries is best for performance

Related Question