Microsoft SQL Server 2008 R2 Unleashed- P138 potx

When you execute the stored procedure with the Show Execution Plan option enabled, SSMS displays only the execution plans for the path or statements that are actually executed, as shown

Stream Aggregate

You most often see the Stream Aggregate operation when you are aggregating a single input,

such as a DISTINCT clause or a SUM, COUNT, MAX, MIN, or AVG operator The output of this

oper-ator may be referenced by later operoper-ators in the query, returned to the client, or both

Because the Stream Aggregate operator requires input ordered by the columns within its

groups, a Sort operator often precedes the Stream Aggregate operator unless the data is

already sorted due to a prior Sort operator or due to an ordered index seek or scan

Table Delete , Table Insert , Table Scan , and Table Update

You see the Table Delete, Table Insert, Table Scan, and Table Update operators when the

indicated operation is being performed against that table as a whole The presence of these

operators does not always mean a problem exists, although a table scan can be an

indica-tor that you might need some indexes to support the query A table scan may be

performed on small tables even if appropriate indexes exist, especially when the table is

only a single page or two in size

Table-valued Function

The Table-valued Function operator is displayed for queries with calls to table-valued

func-tions The Table-valued Function operator evaluates the table-valued function, and the

resulting rows are stored in the tempdb database When the parent operators request the

rows, the Table-valued Function operator returns the rows from tempdb

Top

The Top operator indicates a limit that is set, either by number of rows or a percentage, on

the number of results to be returned from the input The ToolTip may also contain a list

of the columns being checked for ties if the WITH TIES option has been specified

Parallelism Operators

The Parallelism operators indicate that parallel query processing is being performed The

associated logical operator displayed is one of the Distribute Streams, Gather Streams, or

Repartition Streams logical operators

Distribute Streams The Distribute Streams operator takes a single input stream of

records and produces multiple output streams Each record from the input stream

appears in one of the output streams Hashing is typically used to decide to which output

stream a particular input record belongs

Gather Streams The Gather Streams operator consumes several input streams and

produces a single output stream of records by combining the input streams If the output is

ordered, the ToolTip will contain an Order By item indicating the columns being ordered

1315 Query Analysis in SSMS

NOTE

Parallel query processing strategies are covered in more detail in Chapter 35

Analyzing Stored Procedures

When displaying the estimated execution plan for a stored procedure, you see multiple

statement operators as inputs to the Stored Procedure operator, especially if you have any

conditional branching in the stored procedure One operator exists for each statement

defined in the stored procedure When conditional branching occurs in the stored

proce-dure, SQL Server does not know at query optimization time which statements in the

stored procedure will actually be executed, so it has to estimate a query plan for each

indi-vidual statement An example is shown in Figure 36.7

When you execute the stored procedure with the Show Execution Plan option enabled,

SSMS displays only the execution plans for the path or statements that are actually

executed, as shown in Figure 36.8

FIGURE 36.7 Estimated execution plan for a stored procedure

In addition, because stored procedures can become quite complex, with multiple SQL

statements, seeing the graphical execution plan in the SSMS Execution Plan window can

be difficult You might find it easier to break up the stored procedure into smaller batches

or individual queries and analyze it a bit at a time

Saving and Viewing Graphical Execution Plans

SQL Server Management Studio 2008 enables you to save an execution plan as an XML

file To save a graphical execution plan in SSMS, you right-click anywhere on the graphical

execution plan and choose Save Execution Plan As to bring up the Save As dialog

(alterna-tively, you can choose the Save Execution Plan As option from the File menu)

When you save the execution plan to a file, the graphical execution plan is saved as an

XML, file with the sqlplan file extension To view a saved execution plan, click on the

File menu; select Open and then File In the Open File dialog, select Execution Plan files

in the Files of Type drop-down to limit the files displayed to just Execution Plan Files (see

FIGURE 36.8 Actual execution plan used for a stored procedure

1317 Query Analysis in SSMS

tion plan was originally generated, you can mouse over the operators and display the

detailed information contained in the ToolTips

Displaying Execution Plan XML

In addition to viewing the graphical execution plan in SSMS, you can also display the

XML generated by the Query Optimizer that is used to create the graphical execution

plan Right-click on the execution plan and select the Show Execution Plan XML option

(see Figure 36.10)

Selecting this option opens a new XML editor window with the SHOWPLAN_XML output

generated by the query optimizer

Missing Index Hints

One new feature in SQL Server Management Studio 2008 is Missing Index Hints when

displaying the execution plan of a query You can use the Missing Index Hints feature to

help identify columns on which adding an index might help the query execute faster and

more efficiently Missing Index Hints is a lightweight, server-side, always-on feature using

dynamic management objects and execution plans to provide information about missing

indexes that could enhance query performance

FIGURE 36.9 Loading an execution plan into SSMS

FIGURE 36.10 Generating execution plan XML in SSMS

NOTE

The Missing Index Hints feature is separate from the Database Engine Tuning Advisor

available in SQL Server 2008 The Database Engine Tuning Advisor is a more

compre-hensive tool that assesses the physical database design and recommends new

physi-cal design structures for performance improvement In addition to index

recommendations, it also considers whether indexed views or partitioning could be

used to improve query performance

When the query optimizer generates an execution plan, it analyzes what are the best

avail-able indexes for a the specified search and join conditions If a useful index is not found,

the query optimizer generates a suboptimal query plan but still stores information about

the missing indexes The Missing Index Hints feature enables you to view information

about these indexes so you can decide whether they should be implemented

If any missing indexes are identified by the Query Optimizer, the Execution Plan tab in

SSMS displays information related to all the missing indexes If you put the mouse pointer

over the missing index text, it displays a ToolTip showing the T-SQL code required to

create the suggested missing index as suggested, as shown in Figure 36.11

In addition to displaying a ToolTip with the T-SQL code, you can also generate the SQL

1319 Query Analysis in SSMS

FIGURE 36.11 Displaying missing indexes in SSMS

FIGURE 36.12 Generating T-SQL code to create a missing index

LISTING 36.1 SQL Generated by SSMS Missing Index Hints Feature

/*

Missing Index Details from multi_query.sqlplan

The Query Processor estimates that implementing the following index could improve

the query cost by 87.4584%

*/

/*

USE [bigpubs2008]

GO

CREATE NONCLUSTERED INDEX [<Name of Missing Index, sysname,>]

ON [dbo].[sales] ([qty])

INCLUDE ([stor_id])

GO

*/

If you examine the SQL generated by SSMS, notice that it displays the estimated associated

cost benefit expected by adding the recommended index Also note that the script does not

include an index name You need to specify an index name based on your naming standards

NOTE

If you decide to create a recommended index, be sure to review the subsequent query

plan to determine if the query is using the index and that it provides the expected

per-formance benefit

Missing Index Dynamic Management Objects

The Missing Index Hints feature in SSMS draws information regarding missing indexes

from a set of new dynamic management objects introduced in SQL Server 2008:

sys.dm_db_missing_index_group_stats—Returns summary information

about missing index groups, such as the performance improvements that could be

gained by implementing a specific group of missing indexes

sys.dm_db_missing_index_groups—Returns information about a specific group

of missing indexes, such as the group identifier and the identifiers of all missing

indexes contained in that group

sys.dm_db_missing_index_columns—Returns detailed information about a missing

index; for example, it returns the name and identifier of the table where the index is

missing, and the columns and column types that should make up the missing index

sys.dm_db_missing_index_details—Returns information about the database

1321 Query Analysis in SSMS

come in handy After running a typical workload on SQL Server, you can retrieve

informa-tion about missing indexes by querying the dynamic management funcinforma-tions directly You

can use the information returned by these dynamic management objects in scripts and

use the information to generate CREATE INDEX statements to create the missing indexes

Listing 36.2 provides a sample query that displays the missing index information for a

query on the sales table that was run between 10:30 and 10:40 p.m on February 21,

2010

LISTING 36.2 Querying the Missing Index Dynamic Management Objects

SELECT

mig.index_group_handle as handle,

convert(varchar(30), statement) AS table_name,

convert(varchar(12), column_name) AS Column_name,

convert(varchar(10), column_usage) as ColumnUsage,

avg_user_impact as avg_impact

FROM sys.dm_db_missing_index_details AS mid

CROSS APPLY sys.dm_db_missing_index_columns (mid.index_handle)

INNER JOIN sys.dm_db_missing_index_groups AS mig

ON mig.index_handle = mid.index_handle

inner join sys.dm_db_missing_index_group_stats AS migs

ON migs.group_handle = mig.index_group_handle

where mid.object_id = object_id(‘sales’)

and last_user_seek between ‘2010-02-21 22:30’ and ‘2010-02-21 22:40’

ORDER BY mig.index_group_handle, mig.index_handle, column_id;

GO

handle table_name Column_name ColumnUsage avg_impact

- -

-2 [bigpubs -2008].[dbo].[sales] stor_id INCLUDE 87.46

2 [bigpubs2008].[dbo].[sales] qty INEQUALITY 87.46

If you view the output of this query, you see that the optimizer is recommending an index

on the qty column to support an inequality operator It is also recommended that the

stor_id column be specified as an included column in the index This index is estimated

to improve performance by 87.46% When you use this information, the CREATE INDEX

statement for the recommended index would be the following:

CREATE INDEX qty_idx on [bigpubs2008].[dbo].[sales] (qty) INCLUDE (stor_id)

Missing Index Hints Features Limitations

The Missing Index Hints feature provides some helpful information for identifying

poten-tially missing indexes in your database, but it does have a few limitations:

It is not intended to fine-tune the existing indexes, only to recommend additional

indexes when no useful index is found that can be used to satisfy a search or join

condition

It does not specify the order for columns to be specified in the index

For queries involving only inequality predicates, the cost information returned is less

accurate than for equality operators

It only recommends adding included columns to indexes for some queries instead of

creating composite indexes You need to determine whether the included columns

should be specified as additional index key columns instead

It returns only raw information about columns on which indexes might be missing

It may return different costs for the same missing index group for different executions

It does not suggest filtered indexes

The dynamic management objects can store information from a maximum of 500

missing indexes

It is unable to provide recommendations for clustered indexes, indexed views, or

table partitioning (Use the Database Engine Tuning Advisor instead for these

recommendations.)

After the SQL Server is restarted, all the information related to missing indexes is

lost To keep the information for later use, the DBA needs to back up all the data

available within all the missing index dynamic management objects prior to

restart-ing SQL Server

NOTE

Although the Missing Index Hints feature is helpful for identifying indexes that may be

useful for you to define, it’s not a substitute for a well-thought-out index design For

more information on index design, see Chapter 34, “Data Structures, Indexes, and

Performance.”

SSMS Client Statistics

You can use SSMS to get some additional information related to the client-side

perfor-mance of the query by toggling the Include Client Statistics option in the Query menu

When turned on, the Client Statistics tab is added to the Results panel This tab displays

useful performance statistics in a tabular format that is related to how much work the

1323 SSMS Client Statistics

average statistics across all executions Figure 36.13 shows an example of the client

statis-tics displayed after three separate query executions

The first line in the Client Statistics tab displays the actual time the query was executed

The Time Statistics values are specified in number of milliseconds Some of the most

useful pieces of information include the number of rows returned by SELECT statements,

total client processing time, total execution time, and number of bytes sent and received

across the network

The Average column contains the cumulative average since the Include Client Statistics

option was enabled Turning the option off and back on clears out all the historical

statis-tics and resets the averages Alternatively, you can also reset the client statisstatis-tics by

select-ing the Reset Client Statistics option from the Query menu

One of the most helpful features of the client statistics is the arrow indicators provided for

the different executions, which makes it easy to identify which values increased,

decreased, or stayed this same This feature makes it easy to compare the runtime statistics

between different queries or different executions of the same query

FIGURE 36.13 SSMS client statistics