Title: Query Processing and Optimization Dr. Muhammad Shafique
1Query Processing and Optimization Dr. Muhammad
Shafique
2Outline
- Background review
- Processing a query
- SQL queries and relational algebra
- Implementing basic query operations
- Heuristics-based query optimization
- Cost-function-based query optimization
- Semantic-based query optimization
- Query optimization in Oracle DBMS
3Processing a Query
- Query processing and query optimization
- Problem formulation
- Given a query q, a space of execution plans, E,
and a cost function, cost(p) that assigns a
numeric cost to an execution plan p ? E, find the
minimum cost execution plan that computes q.
4Processing a Query
- Typical steps in processing a high-level query
- Query in a high-level query language like SQL
- Scanning, parsing, and validation
- Intermediate-form of query like query tree
- Query optimizer
- Execution plan
- Query code generator
- Object-code for the query
- Run-time database processor
- Results of query
5SQL Queries and Relational Algebra
- SQL query is translated into an equivalent
extended relational algebra expression ---
represented as a query tree - In order to transform a given query into a query
tree, the query is decomposed into query blocks - A query block contains a single
SELECT-FROM-WHERE expression along with GROUP-BY
and HAVING clauses. - The query optimizer chooses an execution plan for
each block
6(No Transcript)
7(No Transcript)
8SQL Queries and Relational Algebra
- Example
- SELECT Lname, Fname
- FROM EMPLOYEE
- WHERE Salary gt ( SELECT MAX(Salary)
- FROM EMPLOYEE
- WHERE Dno 5 )
-
9Translating SQL Queries into Relational Algebra
(1)
- SELECT LNAME, FNAME
- FROM EMPLOYEE
- WHERE SALARY gt ( SELECT MAX (SALARY)
- FROM EMPLOYEE
- WHERE DNO 5)
SELECT MAX (SALARY) FROM EMPLOYEE WHERE DNO 5
SELECT LNAME, FNAME FROM EMPLOYEE WHERE
SALARY gt C
pLNAME, FNAME (sSALARYgtC(EMPLOYEE))
FMAX SALARY (sDNO5 (EMPLOYEE))
10SQL Queries and Relational Algebra (2)
- Uncorrelated nested queries Vs Correlated nested
queries - Example
- Retrieve the name of each employee who works on
all the projects controlled by department number
5. - SELECT FNAME, LNAME FROM EMPLOYEE WHERE
( (SELECT PNO FROM WORKS_ON
WHERE SSNESSN) CONTAINS
(SELECT PNUMBER FROM PROJECT
WHERE DNUM5) )
11SQL Queries and Relational Algebra (3)
- Example
- For every project located in Stafford,
retrieve the project number, the controlling
department number and the department managers
last name, address and birthdate. - SQL query
- SELECT P.NUMBER,P.DNUM,E.LNAME, E.ADDRESS,
E.BDATE - FROM PROJECT AS P,DEPARTMENT AS D, EMPLOYEE AS E
- WHERE P.DNUMD.DNUMBER AND D.MGRSSNE.SSN
AND P.PLOCATIONSTAFFORD - Relation algebra
- ?PNUMBER, DNUM, LNAME, ADDRESS, BDATE
(((?PLOCATIONSTAFFORD(PROJECT)) DNUMDNUMBER
(DEPARTMENT)) MGRSSNSSN (EMPLOYEE)) -
12SQL Queries and Relational Algebra (4)
13Implementing Basic Query Operations
- An RDBMS must provide implementation(s) for all
the required operations including relational
operators and more - External sorting
- Sort-merge strategy
- Sorting phase
- Number of file blocks (b)
- Number of available buffers (nB)
- Runs --- (b / nB)
- Merging phase --- passes
- Degree of merging --- the number of runs that are
merged together in each pass
14Algorithms for External Sorting (1)
- External sorting
- Refers to sorting algorithms that are suitable
for large files of records stored on disk that do
not fit entirely in main memory, such as most
database files. - Sort-Merge strategy
- Starts by sorting small subfiles (runs) of the
main file and then merges the sorted runs,
creating larger sorted subfiles that are merged
in turn.
15Algorithms for External Sorting (2)
16Algorithms for External Sorting (3)
- Analysis
- Number of file blocks b
- Number of initial runs nR
- Available buffer space nB
- Sorting phase nR ?(b/nB)?
- Degree of merging dM Min (nB-1, nR)
- Number of passes nP ?(logdM(nR))?
- Number of block accesses (2 b) (2 b
(logdM(nR))) - Example done in the class
17Heuristic-Based Query Optimization
- Query tree and query transformations
- General transformation rules for relational
algebra operations
18General Transformation Rules for Relational
Algebra Operations
- Cascade of ? A conjunctive selection condition
can be broken up into a cascade (that is, a
sequence) of individual ? operations ? C1 AND
C2 AND .AND Cn (R) ? C1 (?C2( (?Cn(R))) - Commutativity of ? The ? operation is
commutative ?C1(?C2(R)) ?C2(?C1(R)) - Cascade of ? In a cascade (sequence) of ?
operations, all but the last one can be ignored - Commuting ? with ? If the selection condition
c involves only those attributes A1, ..., An in
the projection list, the two operations can be
commuted - And more
19Heuristic-Based Query Optimization
- Outline of heuristic algebraic optimization
algorithm - Break up SELECT operations with conjunctive
conditions into a cascade of SELECT operations - Using the commutativity of SELECT with other
operations, move each SELECT operation as far
down the query tree as is permitted by the
attributes involved in the select condition - Using commutativity and associativity of binary
operations, rearrange the leaf nodes of the tree - Combine a CARTESIAN PRODUCT operation with a
subsequent SELECT operation in the tree into a
JOIN operation, if the condition represents a
join condition - Using the cascading of PROJECT and the commuting
of PROJECT with other operations, break down and
move lists of projection attributes down the tree
as far as possible by creating new PROJECT
operations as needed - Identify sub-trees that represent groups of
operations that can be executed by a single
algorithm
20Heuristic-Based Query Optimization Example
- Query
- "Find the last names of employees born after
1957 who work on a project named Aquarius." - SQL
- SELECT LNAME
- FROM EMPLOYEE, WORKS_ON, PROJECT
- WHERE PNAMEAquarius AND PNUMBERPNO AND
ESSNSSN AND BDATE.1957-12-31
21(No Transcript)
22(No Transcript)
23(No Transcript)
24(No Transcript)
25(No Transcript)
26Implementing Basic Query Operations
- Combining operations using pipelining
- Temporary files based processing
- Pipelining or stream-based processing
- Example consider the execution of the following
query - ?list of attributes( (? c1(R) ? (? c2 (S))
27Implementing Basic Query Operations
- Estimates of selectivity
- Selectivity is the ratio of the number of tuples
that satisfy the condition to the total number of
tuples in the relation. - SELECT ( ? ) operator implementation
- Linear search
- Binary search
- Using a primary index (or hash key)
- Using primary index to retrieve multiple records
- Using clustering index to retrieve multiple
records - Using a secondary index on an equality comparison
- Conjunctive selection using an individual index
- Conjunctive selection using a composite index
- Conjunctive selection by intersection of record
pointers
28Algorithms for JOIN Operations
- Implementing the JOIN Operation
- Join (EQUIJOIN, NATURAL JOIN)
- twoway join a join on two files
- e.g. R AB S
- multi-way joins joins involving more than two
files. - e.g. R AB S CD T
- Examples
- (OP6) EMPLOYEE DNODNUMBER DEPARTMENT
- (OP7) DEPARTMENT MGRSSNSSN EMPLOYEE
29Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Methods for implementing joins
- J1 Nested-loop join (brute force)
- For each record t in R (outer loop), retrieve
every record s from S (inner loop) and test
whether the two records satisfy the join
condition tA sB. - J2 Single-loop join (Using an access structure to
retrieve the matching records) - If an index (or hash key) exists for one of the
two join attributes say, B of S retrieve each
record t in R, one at a time, and then use the
access structure to retrieve directly all
matching records s from S that satisfy sB
tA.
30Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Methods for implementing joins
- J3 Sort-merge join
- If the records of R and S are physically sorted
(ordered) by value of the join attributes A and
B, respectively, we can implement the join in the
most efficient way possible. - Both files are scanned in order of the join
attributes, matching the records that have the
same values for A and B. - In this method, the records of each file are
scanned only once each for matching with the
other fileunless both A and B are non-key
attributes, in which case the method needs to be
modified slightly.
31Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Methods for implementing joins
- J4 Hash-join
- The records of files R and S are both hashed to
the same hash file, using the same hashing
function on the join attributes A of R and B of S
as hash keys. - A single pass through the file with fewer records
(say, R) hashes its records to the hash file
buckets. - A single pass through the other file (S) then
hashes each of its records to the appropriate
bucket, where the record is combined with all
matching records from R.
32Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Factors affecting JOIN performance
- Available buffer space
- Join selection factor
- Choice of inner VS outer relation
33Buffer Space and Join performance
-
- In the nested-loop join, it makes a difference
which file is chosen for the outer loop and which
for the inner loop. If EMPLOYEE is used for the
outer loop, each block of EMPLOYEE is read once,
and the entire DEPARTMENT file (each of its
blocks) is read once for each time we read in (nB
- 2) blocks of the EMPLOYEE file. We get the
following - Total number of blocks accessed for outer file
bE - Number of times ( nB - 2) blocks of outer file
are loaded ? bE / nB 2 ? - Total number of blocks accessed for inner file
bD ?bE / nB 2 ? - Hence, we get the following total number of
block accesses - bE (? bE / nB 2 ? bD) 2000 (? (2000/5)
? 10) 6000 blocks -
- On the other hand, if we use the DEPARTMENT
records in the outer loop, by symmetry we get the
following total number of block accesses - bD (? bD / nB 2 ? bE) 10 (?(10/5) ?
2000) 4010 blocks -
34Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Other types of JOIN algorithms
- Partition hash join
- Partitioning phase
- Each file (R and S) is first partitioned into M
partitions using a partitioning hash function on
the join attributes - R1 , R2 , R3 , ...... RM and S1 , S2 , S3 ,
...... SM - Minimum number of in-memory buffers needed for
the partitioning phase M1. - A disk sub-file is created per partition to store
the tuples for that partition. - Joining or probing phase
- Involves M iterations, one per partitioned file.
- Iteration i involves joining partitions Ri and
Si.
35Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Partitioned Hash Join Procedure
- Assume Ri is smaller than Si.
- Copy records from Ri into memory buffers.
- Read all blocks from Si, one at a time and each
record from Si is used to probe for a matching
record(s) from partition Si. - Write matching record from Ri after joining to
the record from Si into the result file.
36Algorithms for JOIN Operations
- Implementing the JOIN Operation (contd.)
- Cost analysis of partition hash join
- Reading and writing each record from R and S
during the partitioning phase (bR bS),
(bR bS) - Reading each record during the joining
phase (bR bS) - Writing the result of join bRES
- Total Cost
- 3 (bR bS) bRES
37Query Optimization Using Selectivity and Cost
Function
- Cost-based query optimization
- Estimate and compare the costs of executing a
query using different execution strategies and
choose the strategy with the lowest cost
estimate. - Issues
- Cost function
- Number of execution strategies to be considered
- Compare to heuristic query optimization
38Query Optimization Using Selectivity and Cost
Function
- Cost Components for Query Execution
- Access cost to secondary storage
- Storage cost
- Computation cost
- Memory usage cost
- Communication cost
- Note Different database systems may focus on
different cost components.
39Query Optimization Using Selectivity and Cost
Function
- Catalog Information Used in Cost Functions
- Information about the size of a file
- number of records (tuples) (r),
- record size (R),
- number of blocks (b)
- blocking factor (bfr)
- Information about indexes and indexing attributes
of a file - Number of levels (x) of each multilevel index
- Number of first-level index blocks (bI1)
- Number of distinct values (d) of an attribute
- Selectivity (sl) of an attribute
- Selection cardinality (s) of an attribute. (s
sl r)
40Query Optimization Using Selectivity and Cost
Function
- Examples of Cost Functions for SELECT
- S1. Linear search (brute force) approach
- CS1a b
- For an equality condition on a key, CS1a (b/2)
if the record is found otherwise CS1a b. - S2. Binary search
- CS2 log2b (s/bfr)? 1
- For an equality condition on a unique (key)
attribute, CS2 log2b - S3. Using a primary index (S3a) or hash key (S3b)
to retrieve a single record - CS3a x 1 CS3b 1 for static or linear
hashing - CS3b 2 for extendible hashing
41Query Optimization Using Selectivity and Cost
Function
- Examples of Cost Functions for SELECT (contd.)
- S4. Using an ordering index to retrieve multiple
records - For the comparison condition on a key field with
an ordering index, CS4 x (b/2) - S5. Using a clustering index to retrieve multiple
records - CS5 x (s/bfr)
- S6. Using a secondary (B-tree) index
- For an equality comparison, CS6a x s
- For a comparison condition such as gt, lt, gt, or
lt, - CS6a x (bI1/2) (r/2)
42Query Optimization Using Selectivity and Cost
Function
- Examples of Cost Functions for SELECT (contd.)
- S7. Conjunctive selection
- Use either S1 or one of the methods S2 to S6 to
solve. - For the latter case, use one condition to
retrieve the records and then check in the memory
buffer whether each retrieved record satisfies
the remaining conditions in the conjunction. - S8. Conjunctive selection using a composite
index - Same as S3a, S5 or S6a, depending on the type of
index. - Examples of using the cost functions.
43Query Optimization Using Selectivity and Cost
Function
- Examples of Cost Functions for JOIN
- Join selectivity (js)
- js (R C S) / R x S (R C S)
/ (R S ) - If condition C does not exist, js 1
- If no tuples from the relations satisfy condition
C, js 0 - Usually, 0 lt js lt 1
- Size of the result file after join operation
- (R C S) js R S
44Query Optimization Using Selectivity and Cost
Function
- Examples of Cost Functions for JOIN (contd.)
- J1. Nested-loop join
- CJ1 bR (bRbS) ((js R S)/bfrRS)
- (Use R for outer loop)
- J2. Single-loop join (using an access structure
to retrieve the matching record(s)) - If an index exists for the join attribute B of S
with index levels xB, we can retrieve each record
s in R and then use the index to retrieve all the
matching records t from S that satisfy tB
sA. - The cost depends on the type of index.
45Query Optimization Using Selectivity and Cost
Function
- Examples of Cost Functions for JOIN (contd.)
- J2. Single-loop join (contd.)
- For a secondary index,
- CJ2a bR (R (xB sB)) ((js R
S)/bfrRS) - For a clustering index,
- CJ2b bR (R (xB (sB/bfrB))) ((js
R S)/bfrRS) - For a primary index,
- CJ2c bR (R (xB 1)) ((js R
S)/bfrRS) - If a hash key exists for one of the two join
attributes B of S - CJ2d bR (R h) ((js R S)/bfrRS)
- J3. Sort-merge join
- CJ3a CS bR bS ((js R S)/bfrRS)
- (CS Cost for sorting files)
46Query Optimization Using Selectivity and Cost
Function
- Multiple Relation Queries and Join Ordering
- A query joining n relations will have n-1 join
operations, and hence can have a large number of
different join orders when we apply the algebraic
transformation rules. - Current query optimizers typically limit the
structure of a (join) query tree to that of
left-deep (or right-deep) trees. - Left-deep tree
- A binary tree where the right child of each
non-leaf node is always a base relation. - Amenable to pipelining
- Could utilize any access paths on the base
relation (the right child) when executing the
join.
47Query Optimization Using Selectivity and Cost
Function
48Cost-Function Based Query Optimization
- Example
- SELECT pnumber, dnum, lname, address, bdate
- FROM Project, Department, Employee
- WHERE dnumdnumber AND mgrssnssn AND plocation
Dhahran - Join order selection
- Project Department Employee
- Department Project Employee
- Department Employee Project
- Employee Department Project
49Cost-Function Based Query Optimization
50Traditional Query optimizers
- Three design components of commercial query
optimizers - Execution space
- Normally represented as an annotated query tree
- All trees that compute the query are considered
legal plans for the query - A finite subset of the infinite space of legal
plans is the search space - Cost model
- It assigns an integer cost to a plan based on
some assumptions about the statistical
distribution of data and the abstract machine - Search algorithm
- Algorithm used to search the search space for the
plan with minimum cost. For example, search
algorithm for System R is dynamic programming
51A Model for Parallel Execution
- Two possible sources of parallelism
- Inter-operator parallelism
- Data independent like sub-trees and data
dependent like pipelined - Intra-operator parallelism
- Like sort
- Two deterrents of parallelism
- Data dependencies between operators
- Resource contention
- Several operators running in parallel and
competing for the same resources - Operator tree
- Operator tree has nodes that are atomic, i.e.,
run-time scheduler cannot subdivide the operation
further
52Dynamic Programming Algorithm
- DP Algorithm for left-Deep join trees
- Input An SPJ query q on relations R1, R2, , Rn
- Output A query plan Popt for q with optimal work
- for i 1 to n do
- optPlan(Ri) accessPlan(Ri)
- for i 2 to n do
- for all S ? R1, R2, , Rn s.t. S i do
- bestPlan dummy plan with infinite cost
- for all Rj, Sj s.t. S Rj ? Sj do
- p joinPlan(optPlan(Sj) , Rj)
- if p ? work bestPlan then bestPlan p
- optPlan(S) bestPlan
- Popt optPlan( R1, R2, ., Rn )
- Principle of optimality An optimal plan for a
set of relations is an extension of an optimal
plan for some subset of the set
53Failure of Dynamic Programming
- Fundamental assumptions in DP
- Principle of optimality
- p1 cost p2 ? (?i) joinplan(p1, Ri) cost
joinplan(p2, Ri) - Total order
- (?p1p2) not(p1 cost p2) ? (p2 cost p1)
- DP and response time
- Any reasonable cost model for response time will
violate the principle of optimality - Generalization of DP for partial orders
54Overview of Query Optimization in Oracle
- Rule-based query optimization the optimizer
chooses execution plans based on heuristically
ranked operations. - May be phased out
- Cost-based query optimization the optimizer
examines alternative access paths and operator
algorithms and chooses the execution plan with
lowest estimate cost. - The query cost is calculated based on the
estimated usage of resources such as I/O, CPU and
memory needed. - Application developers could specify hints to the
ORACLE query optimizer. - application developer might know more information
about the data. - SELECT / ...hint... / rest of query
- SELECT / index(t1 t1_abc) index(t2 t2_abc) /
COUNT()FROM t1, t2WHERE t1.col1 t2.col1
55Query Processing and Optimization in Oracle RDB
- Applications using
- Embedded SQL
- SQL module language
- Dynamic SQL
- Interactive SQL
- Distributed and heterogeneous query processing
- DEC DB Integrator (also known as Access Works)
- Sub-queries can be submitted to other DBMSs (DB2,
Sybase - DEC DB Integrator has its own query optimizer
56Query Processing and Optimization in Oracle RDB
- Oracle query processor designed for wide range of
queries from simple to very complex - Queries with up to 128 joins can be evaluated
- Query processor consists of two parts
- Query compiler
- Query executor
- Each part has its own optimizer
- Query transformation
- Takes place in the SQL front-end
- SQL text to polish notation to linked structure
of internal blocks - Other transformations take place during the other
stages like code generation
57Query Processing and Optimization in Oracle RDB
- Execution plan selection
- To determine a join order
- Exhaustive search
- Left-deep tree of two-way joins for each
sub-query - The least expensive join order for each sub-query
starting from innermost - Execution plan selection based on the cost model
- Join strategies
- Supports nested-loop join
- Two-way merge join works on two sorted streams
58Query Processing and Optimization in Oracle RDB
- Single table scan
- Sequential scan
- Scan of one or more indexes
- Direct access of a row by the row identifier
(RID) - Dynamic optimization
- Advanced features
- BLOBs
- Triggers
- NOT NULL and UNIQUE constraints are not evaluated
when a row is deleted from a table
59Semantic Query Optimization
- Semantic query optimization uses constraints
specified in the database schema - The technique may be used in combination with
other techniques - Example
- SELECT E.LNAME, M.LNAME
- FROM EMPLOYEE AS E, EMPLOYEE AS M
- WHERE E.SUPERSSNM.SSN AND E.SALARY gtM.SALARY
- Constraint --- no employee can earn more than his
manager - No need to execute the query
- Searching through the constraints to find out
which one is applicable !!!
60References
- Required
- Textbook Chapter 15
- Execution Strategies for SQL Subqueries
Mostafa Elhamali et al SIGMOD07 June 12-14,
2007, Beijing China pp 993-1004 - Recommended
- Adaptive Query Processing Why, How, When, What
next Amol Deshpande et al, SIGMOD 2006, pp
806-807. - Efficient and Extensible Algorithms for Multi
Query OptimizationsPrasan Rop, S. Seshadri, S.
Sudarshan, and S. bhobeACM SIGMOD Record volume
29, Number 2, June 2000, pp 249 260 - Query Processing and Optimization in Oracle
RDBGennady Antoshenkov and Mohamed ZiauddinThe
VLDB Journal Springer-Verlag, VLDB(1996) 5 229
237
61Summary
- Background review
- Processing a query
- SQL queries and relational algebra
- Implementing basic query operations
- Heuristics-based query optimization
- Cost-function-based query optimization
- Query Processing and Optimization in Oracle RDB