understanding graphical execution plans level 200

29
Ashwani Roy Understanding Graphical Execution Plans Level 200

Upload: nadine

Post on 24-Feb-2016

43 views

Category:

Documents


0 download

DESCRIPTION

Understanding Graphical Execution Plans Level 200. Ashwani Roy . Agenda. Query Processing lifecycle by Database Engine Elements in a Execution Plans Important Execution Plan Operators. What Happens when a Query is submitted. Query Parsing Query Optimization Query Execution . - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Understanding Graphical Execution Plans Level 200

Ashwani Roy

Understanding Graphical Execution Plans Level 200

Page 2: Understanding Graphical Execution Plans Level 200

• Query Processing lifecycle by Database Engine

• Elements in a Execution Plans

• Important Execution Plan Operators

Agenda

Page 3: Understanding Graphical Execution Plans Level 200

• Query Parsing

• Query Optimization

• Query Execution

What Happens when a Query is submitted

Page 4: Understanding Graphical Execution Plans Level 200

Query Plan

Page 5: Understanding Graphical Execution Plans Level 200

• Logical and Physical Operators • Parallelism Physical Operators• Cursor Operators• Language Elements

Operators in an Execution Plan

Page 6: Understanding Graphical Execution Plans Level 200

Columns in a Plan

Rows EstimateIO

Executes EstimateCPU

StmtId AvgRowSize

NodeId TotalSubtreeCost

Parent OutputList

PhysicalOp Warnings

LogicalOp Type

Argument Parallel

DefinedValues EstimateExecutions

EstimateRows

Page 7: Understanding Graphical Execution Plans Level 200

• If a Cached Plan exists then SQL Server will use this cached plan

DEMO 01

Cached Query Plans

Page 8: Understanding Graphical Execution Plans Level 200

Important Operators in Execution Plans

Select (Result)

Sort   Clustered Index Seek  

Clustered Index Scan    

Non-clustered Index Scan

Non-clustered Index Seek  

Table Scan RID Lookup Key Lookup Hash Match  

Nested Loops  

Merge Join Top Compute Scalar

Constant Scan

Filter Lazy Spool Spool Eager Spool Stream Aggregate

Distribute Streams 

Repartition Streams

Gather Streams 

Bitmap Split

Page 9: Understanding Graphical Execution Plans Level 200

Index Seek

• Reads B-tree entries to determine the data page• The Argument column contains the name of the

nonclustered index being used• Prefered for highly selective queries

Page 10: Understanding Graphical Execution Plans Level 200

Index Seek

Page 11: Understanding Graphical Execution Plans Level 200

Index Scan

• Horizontal traversal of the leaf level of the index from the first page to the last

• Retrieves all rows from the nonclustered index• The Argument column contains the name of the

nonclustered index being used

Page 12: Understanding Graphical Execution Plans Level 200

Clustered Index Scan

• The clustered index scan’s logical and physical operator scans the clustered index

• The Argument column contains the name of the clustered index

• If the table does not have Clustered Index the same Query will produce Table Scan

Page 13: Understanding Graphical Execution Plans Level 200

DEMO 02

Page 14: Understanding Graphical Execution Plans Level 200

Clustered Index Seek

• Cluster index seek • Uses the seeking ability of indexes to retrieve rows• The Argument column contains the name of the

clustered index being used• Seek() predicate contains the columns used for

seeking

Page 15: Understanding Graphical Execution Plans Level 200

Bookmark Lookups

• Uses a bookmark to look up a row in a clustered index or table

• The Argument column contains the bookmark label

• Can be removed by covering columns• May have a performance improvement

Page 16: Understanding Graphical Execution Plans Level 200

KEY LOOKUP

• A Key Lookup is a bookmark lookup on a table with a clustered index.

• Means that the optimizer cannot retrieve the rows in a single operation, and has to use a clustered key (or a row ID) to return the corresponding rows from a clustered index (or from the table itself).

• Performance can be improved by making Non-Clustered Index or Covering Index

Page 17: Understanding Graphical Execution Plans Level 200

DEMO 03

Page 18: Understanding Graphical Execution Plans Level 200

RID Lookup

• A type of bookmark lookup

• Occurs on a heap table (a table that doesn't have a clustered index)

• Uses a row identifier to find the rows to return.

Page 19: Understanding Graphical Execution Plans Level 200

DEMO 04

Page 20: Understanding Graphical Execution Plans Level 200

Nested Loop

• The top input to the nested loop is the outer table• The bottom input to the nested loop is the inner table• For each outer row, searches for matching rows are in

the inner input table• Effective if the outer input is very small and the inner

input is preindexed and very large• Optimizer sometimes sorts the outer input to improve

locality of the searches on the index over the inner input

• Best when search exploits an index (indexes on join columns are used)

• Low memory requirement

Page 21: Understanding Graphical Execution Plans Level 200

Hash Join

• The top input is build input, the smaller of the two inputs

• The bottom input is probe input• The hash join first scans or computes the whole

build input• Requires at least one equality clause in the join

predicate• Good for ad-hoc queries

Page 22: Understanding Graphical Execution Plans Level 200

Merge Join

• Both inputs should be sorted on the merge column keys

• An index on a correct set of columns is useful• A many-to-many merge join uses a temporary

table to store rows• Very fast if the data that you want can be

obtained presorted from existing B-tree indexes

Page 23: Understanding Graphical Execution Plans Level 200

WHICH JOIN IS GOOD

• NONE AND ALL• A Merge Join is an efficient way to join two tables,#

• when the join columns are pre sorted • if the join columns are not pre sorted, the query

optimizer has the option of• a) sorting the join columns first, then performing

a Merge Join, or • b) performing a less efficient Hash Join. The

query optimizer considers all the options and generally chooses the execution plan that uses the least resources.

Page 24: Understanding Graphical Execution Plans Level 200

DEMO 05

Page 25: Understanding Graphical Execution Plans Level 200

Stream Aggregation

• The argument column of the plan output shows the list of columns of the GROUP BY or DISTINCT clause

• The list of aggregate expressions will appear in the Defined Values column of the plan output

• Best for smaller sets or sets already sorted• Input is sorted and output is ordered

Page 26: Understanding Graphical Execution Plans Level 200

Hash Aggregation

• Used with large sets• Aggregations are evaluated while building the

hash• Input can be in random order; output is always

in random order

Page 27: Understanding Graphical Execution Plans Level 200

DEMO 06

Page 28: Understanding Graphical Execution Plans Level 200

Rewinds and Rewinds

Page 29: Understanding Graphical Execution Plans Level 200

Questions THANK YOU