Relational Query

Optimization
CS186
R & G Chapters 12/15

Review
Implementation of single Relational Operations
Choices depend on indexes, memory, stats,
Joins
Blocked nested loops:
simple, exploits extra memory

Indexed nested loops:

best if 1 rel small and one indexed

Sort/Merge Join
good with small amount of memory, bad with duplicates

Hash Join
fast (enough memory), bad with skewed data

Query Optimization Overview

Query can be converted to relational algebra

Rel. Algebra converted to tree, joins as branches
Each operator has implementation choices
Operators can also be applied in different order!

SELECT S.sname
FROM Reserves R,

Sailors

S
R.sid=S.sid AND
R.bid=100 AND
S.rating>5
(sname)(bid=100 rating > 5) (Reserves Sailors)
WHERE

sname

bid=100

rating > 5

sid=sid

Reserves

Sailors

Query Optimization Overview

(cont.)
Plan: Tree of R.A. ops (and some others)
with choice of algorithm for each op.

Recall: Iterator interface (next()!)

Three main issues:

For a given query, what plans are considered?

How is the cost of a plan estimated?

How do we search in the plan space?

Ideally: Want to find best plan.

Reality: Avoid worst plans!

Issue 1: Plan Space

Programs that compute
query

Plans that work with

our choice of operators

Plans that we may

compute the cost of

Issue 2: Cost a Plan

sname

(sort project)

bid=100

rating > 5

(inline selection)

sid=sid

Reserves

(hash join)

Sailors

This will cost

about 6472 disk
acccess, given
what I know
about the
database!

Issue 3: Plan Search (AI 101!)

Plans that work with
our choice of operators

Plans that we may

compute the cost of

Cost-based Query SubSystem

Queries

Select *
From Blah B
Where B.blah = blah

Query Parser

Usually there is a
heuristics-based
rewriting step before
the cost-based steps.

Query Optimizer
Plan
Generator

Plan Cost
Estimator

Query Executor

Catalog Manager

Schem
a

Statistic
s

Schema for Examples

Sailors (sid: integer, sname: string, rating: integer, age

Reserves (sid: integer, bid: integer, day: dates, rname:
As seen in previous lectures
Reserves:
Each tuple is 40 bytes long, 100 tuples per page,
1000 pages.
Assume there are 100 boats
Sailors:
Each tuple is 50 bytes long, 80 tuples per page, 500
pages.
Assume there are 10 different ratings
Assume we have 5 pages in our buffer pool!

Motivating Example
SELECT S.sname
FROM Reserves R,

Sailors

S
R.sid=S.sid AND
R.bid=100 AND
Cost: 500+500*1000 I/Os
S.rating>5
Plan:
WHERE

sname

(On-the-fly)

By no means the worst plan!

Misses several opportunities:
rating > 5 (On-the-fly)
selections could have been
bid=100
`pushed earlier, no use is
made of any available indexes,
etc.
(Page-Oriented
sid=sid Nested loops)
Goal of optimization: To find
more efficient plans that
compute the same answer.
Reserves
Sailors

Page-Oriented Nested Loops Join

< S:foreach page bR in R do
R>
foreach page bS in S do
foreach tuple r in bR do
foreach tuple s in bSdo
if ri == sj then add <r, s> to result
Cost = [R]*[S] + [R] = 1000*500 + 1000
If smaller relation (S) is outer, cost = 500*1000 + 500
Much better than nave per-tuple approach!

Alternative Plans Push Selects

(No Indexes)

(On-the-fly)

sname

(On-the-fly)

sname

bid=100
bid=100

rating > 5

(On-the-fly)

(On-the-fly)
(Page-Oriented
sid=sid Nested loops)

(Page-Oriented
sid=sid Nested loops)

Sailors

Reserves

500,500 IOs

rating > 5

(On-the-fly) Reserves

Sailors

250,500 IOs

Alternative Plans Push Selects

(No Indexes)
sname

(On-the-fly)
sname

bid=100

(On-the-fly)

(Page-Oriented
sid=sid Nested loops)
rating > 5

(On-the-fly) Reserves

(On-the-fly)

(Page-Oriented
sid=sid Nested loops)
bid = 100

rating > 5

(On-the-fly)
Sailors

(On-the-fly)
Reserves

Sailors

250,500 IOs

250,500 IOs

Alternative Plans Push Selects

(No Indexes)
sname

bid=100

(On-the-fly)

rating > 5

(On-the-fly)

(On-the-fly) Reserves

Sailors

250,500 IOs

(On-the-fly)

(Page-Oriented
sid=sid Nested loops)

(Page-Oriented
sid=sid Nested loops)
rating > 5

(On-the-fly)

sname

bid=100

(On-the-fly)

Sailors

Reserves

6000 IOs

Alternative Plans Push Selects

(No Indexes)
sname

(On-the-fly)
sname

rating > 5

(On-the-fly)
(Page-Oriented
sid=sid Nested loops)

(Page-Oriented
sid=sid Nested loops)
bid=100
bid=100

(On-the-fly)

Sailors

(On-the-fly)
Reserves

Reserves

6000 IOs

(On-the-fly)

(Scan &
to
temp T2)

rating > 5 Write

Sailors

4250 IOs
1000 + 500+ 250 + (10 * 250)

Alternative Plans Push Selects

(No Indexes)
(On-the-fly)

sname

sname

(Page-Oriented
sid=sid Nested loops)
(Scan &
to
temp T2)

rating > 5 Write

bid=100

(On-the-fly)
Reserves

4250 IOs

Sailors

(On-the-fly)

(Page-Oriented
sid=sid Nested loops)

rating>5

(On-the-fly)
Sailors

bid=100

(Scan &
Write to
temp T2)

Reserves

4010 IOs
500 + 1000 +10 +(250 *10)

(On-the-fly)

More Alternative Plans

(No Indexes)
Main difference:
Sort Merge Join

sname

sid=sid

(Scan;
write to bid=100
temp T1)

(Sort-Merge Join)

rating > 5

(Scan;
write to
temp T2)

Reserves
Sailors
With 5 buffers, cost of plan:
Scan Reserves (1000) + write temp T1 (10 pages, if we
have 100 boats, uniform distribution) = 1010.
Scan Sailors (500) + write temp T2 (250 pages, if have 10 ratings) =
750.
Sort T1 (2*2*10) + sort T2 (2*4*250) + merge (10+250) = 2300

Total: 4060 page I/Os.

If use BNL join, join = 10+4*250, total cost = 2770.
Can also `push projections, but must be careful!
T1 has only sid, T2 only sid, sname:
T1 fits in 3 pgs, cost of BNL under 250 pgs, total < 2000.

More Alt Plans: Indexes

sname

(On-the-fly)

(On-the-fly)
With clustered index on
rating > 5
bid of Reserves, we get
(Index Nested Loops,
100,000/100 = 1000
sid=sid with pipelining )
tuples on 1000/100 = 10 (Use hash
Index, do
pages.
bid=100
Sailors
not write
to temp)
INL with outer not
Reserves
materialized.
Projecting out unnecessary fields from

outer

doesnt help.

Join column sid is a key for Sailors.

At most one matching tuple, unclustered index on sid O

Decision not to push rating>5 before the join is based on

availability of sid index on Sailors.
Cost: Selection of Reserves tuples (10 I/Os); then, for ea
must get matching Sailors tuple (1000*1.2); total 1210 I

What is needed for optimization?

A closed set of operators
Relational ops (table in, table out)
Encapsulation based on iterators
Plan space, based on
Based on relational equivalences, different
implementations
Cost Estimation, based on
Cost formulas
Size estimation, based on
Catalog information on base tables
Selectivity (Reduction Factor) estimation

A search algorithm
To sift through the plan space based on cost!

Summary
Query optimization is an important task in a
relational DBMS.
Must understand optimization in order to
understand the performance impact of a given
database design (relations, indexes) on a
workload (set of queries).
Two parts to optimizing a query:
Consider a set of alternative plans.
Must prune search space; typically, left-deep plans only.

Must estimate cost of each plan that is considered.

Must estimate size of result and cost for each plan node.
Key issues: Statistics, indexes, operator implementations.

Query Optimization
Query can be dramatically improved by
changing access methods, order of operators.
Iterator interface
Cost estimation
Size estimation and reduction factors
Statistics and Catalogs
Relational Algebra Equivalences
Choosing alternate plans
Multiple relation queries
Will focus on System R-style optimizers

Highlights of System R
Optimizer
Impact:
Most widely used currently; works well for < 10 joins.
Cost estimation:
Very inexact, but works ok in practice.
Statistics, maintained in system catalogs, used to estimate cost
of operations and result sizes.
Considers combination of CPU and I/O costs.
More sophisticated techniques known now.
Plan Space: Too large, must be pruned.
Many plans share common, overpriced subtrees

ignore them all!

In some implementations, only the space of left-deep plans is

considered.
Cartesian products avoided in some implementations.

Query Blocks: Units of

Optimization
An SQL query is parsed into a
collection of query blocks, and these
are optimized one block at a time.
Nested blocks are usually treated as
calls to a subroutine, made once per
outer tuple. (This is an oversimplification, wait til we learn more
about nested queries.)

SELECT S.sname
FROM Sailors S
WHERE S.age IN
(SELECT MAX (S2.age
FROM Sailors S2
GROUP BY S2.rating

Outer block Nested bloc

For each block, the plans considered are:
All available access methods, for
D
each relation in FROM clause.
C
All left-deep join trees (i.e., right branch
always a base table, consider
B
A
all join orders and join methods.)

Schema for Examples

Sailors (sid: integer, sname: string, rating: integer, age

Reserves (sid: integer, bid: integer, day: dates, rname:
Reserves:
Each tuple is 40 bytes long, 100 tuples per page,
1000 pages. 100 distinct bids.
Sailors:
Each tuple is 50 bytes long, 80 tuples per page, 500
pages. 10 ratings, 40,000 sids.

Translating SQL to Relational

Algebra
SELECT S.sid, MIN (R.day)
FROM Sailors S, Reserves R, Boats B
WHERE S.sid = R.sid AND R.bid = B.bid AND B.color =
red
GROUP BY S.sid
HAVING COUNT (*) >= 2

For each sailor with at least two reservations for

red boats, find the sailor id and the earliest date on
which the sailor has a reservation for a red boat.

Translating SQL to Relational

Algebra
SELECT S.sid, MIN (R.day)
FROM Sailors S, Reserves R, Boats B
WHERE S.sid = R.sid AND R.bid = B.bid AND B.color =
red
GROUP BY S.sid
HAVING COUNT (*) >= 2

S.sid, MIN(R.day)

(HAVING COUNT(*)>2 (
GROUP BY S.Sid (
B.color = red (
Sailors Reserves

Boats))))

Relational Algebra
Equivalences
Allow us to choose different join orders and to `push
selections and projections ahead of joins.
Selections:

c1cn(R) c1((cn(R))) (cascade)

c1(c2(R)) c2(c1(R))
(commute)
Projections:

a1(R) a1((a1, , an(R))) (cascade)

Cartesian Product

R (S T) (R S) T
(associative)
RS SR
(commutative)
This means we can do joins in any order.
Butbeware of cartesian product!

More Equivalences
Eager projection
Can cascade and push some projections thru
selection
Can cascade and push some projections below one
side of a join
Rule of thumb: can project anything not needed
downstream
Selection between attributes of the two arguments
of a cross-product converts cross-product to a join.
A selection on just attributes of R commutes with
R
S. (i.e., (R
S) (R)
S)

Cost Estimation
For each plan considered, must estimate total
cost:
Must estimate cost of each operation in plan tree.
Depends on input cardinalities.
Weve already discussed how to estimate the cost of
operations (sequential scan, index scan, joins, etc.)

Must estimate size of result for each operation in

tree!
Use information about the input relations.
For selections and joins, assume independence of predicates.

In System R, cost is boiled down to a single number

consisting of #I/O + factor * #CPU instructions
Q: Is cost the same as estimated run time?

Statistics and Catalogs

Need information about the relations and indexes
involved. Catalogs typically contain at least:
# tuples (NTuples) and # pages (NPages) per reln.
# distinct key values (NKeys) for each index.
low/high key values (Low/High) for each index.
Index height (IHeight) for each tree index.
# index pages (INPages) for each index.
Catalogs updated periodically.
Updating whenever data changes is too expensive;
lots of approximation anyway, so slight inconsistency
ok.
More detailed information (e.g., histograms of
the values in some field) are sometimes stored.

Size Estimation and Reduction

Factors
SELECT attribute list
FROM relation list
WHERE term1 AND ... AND

term

Consider a query block:

Maximum # tuples in result is the product of the
cardinalities of relations in the FROM clause.
Reduction factor (RF) associated with each term
reflects the impact of the term in reducing result size.
Result cardinality = Max # tuples * product of all RFs.
RF usually called selectivity
only R&G seem to call it Reduction Factor
beware of confusion between high selectivity as defined
here and highly selective in common English!

Result Size Estimation

Result cardinality =
tuples * product of all RFs.
Term col=value (given index I on col )
RF = 1/NKeys(I)
Term col1=col2 (This is handy for joins too)
RF = 1/MAX(NKeys(I1), NKeys(I2))
Term col>value
RF = (High(I)-value)/(High(I)-Low(I))

Max #

(Implicit assumptions: values are uniformly distributed and terms

are independent!)
Note, if missing indexes, assume 1/10!!!

Postgres 8:
include/utils/selfuncs.h

/* default selectivity estimate

for equalities such as "A = b"
*/
#define DEFAULT_EQ_SEL 0.005
/* default selectivity estimate
for inequalities such as "A <
b" */
#define DEFAULT_INEQ_SEL
0.3333333333333333
/* default selectivity estimate
for range inequalities "A > b
AND A < c" */
#define DEFAULT_RANGE_INEQ_SEL
0.005

/* default selectivity estimate

for pattern-match operators
such as LIKE */
#define DEFAULT_MATCH_SEL
0.005
/* default number of distinct
values in a table */
#define DEFAULT_NUM_DISTINCT 200
/* default selectivity estimate
for boolean and null test
nodes */
#define DEFAULT_UNK_SEL
0.005
#define DEFAULT_NOT_UNK_SEL
(1.0 - DEFAULT_UNK_SEL)

Backend/optimizer/path/clauses
el.c
/*
*
* THIS IS A HACK TO GET V4 OUT THE DOOR.
*
-- JMH 7/9/92
*/
s1 = (Selectivity) 0.3333333;

Think through estimation for

joins

Term col1=col2
RF = 1/MAX(NKeys(I1), NKeys(I2))

Q: Given a join of R and S, what is the range of

possible result sizes (in #of tuples)?
If join is on a key for R (and a Foreign Key in S)?
A common case, can treat it specially

General case: join on {A} ({A} is key for neither)

estimate each tuple r of R generates
NTuples(S)/NKeys(A,S) result tuples, so
NTuples(R) * NTuples(S)/NKeys(A,S)
but can also consider it starting with S, yielding:
NTuples(S) * NTuples(R)/NKeys(A,R)
If these two estimates differ, take the lower one!
Q: Why?

S.{A}

Enumeration of Alternative Plans

There are two main cases:
Single-relation plans
Multiple-relation plans
For queries over a single relation, queries consist of a
combination of selects, projects, and aggregate ops:
Each available access path (file scan / index) is considered,
and the one with the least estimated cost is chosen.
The different operations are essentially carried out
together (e.g., if an index is used for a selection, projection
is done for each retrieved tuple, and the resulting tuples
are pipelined into the aggregate computation).

Cost Estimates for Single-Relation

Plans
Index I on primary key matches selection:
Cost is Height(I)+1 for a B+ tree.
Clustered index I matching one or more selects:

(NPages(I)+NPages(R)) * product of RFs of matching selects.

Non-clustered index I matching one or more

selects:

(NPages(I)+NTuples(R)) * product of RFs of matching selects.

Sequential scan of file:

NPages(R).
Recall: Must also charge for duplicate elimination if required

Example
If we have an index on rating:

SELECT S.sid
FROM Sailors
WHERE

S.rating=8

Cardinality = (1/NKeys(I)) * NTuples(R) = (1/10) * 40000 tuples

Clustered index: (1/NKeys(I)) * (NPages(I)+NPages(R)) = (1/10) *
(50+500) = 55 pages are retrieved. (This is the cost.)
Unclustered index: (1/NKeys(I)) * (NPages(I)+NTuples(R)) = (1/10) *
(50+40000) = 401 pages are retrieved.

If we have an index on sid:

Would have to retrieve all tuples/pages. With a clustered index, the

cost is 50+500, with unclustered index, 50+40000.

Doing a file scan:

We retrieve all file pages (500).

Queries Over Multiple

Relations

A heuristic decision in System R:

only left-deep join trees are considered.

As the number of joins increases, the number of alternative

plans grows rapidly; we need to restrict the search space.
Left-deep trees allow us to generate all fully pipelined plans.
Intermediate results not written to temporary files.
Not all left-deep trees are fully pipelined (e.g., SM join).

D
C
A

C
A

Enumeration of Left-Deep
Plans
Left-deep plans differ only in the order of
relations, the access method for each relation,
and the join method for each join.
Enumerated using N passes (if N relations joined):

Pass 1: Find best 1-relation plan for each relation.

Pass 2: Find best way to join result of each 1-relation plan (as
outer) to another relation. (All 2-relation plans.)
Pass N: Find best way to join result of a (N-1)-relation plan (as
outer) to the Nth relation. (All N-relation plans.)

For each subset of relations, retain only:

Cheapest plan overall, plus

Cheapest plan for each interesting order of the tuples.

The Dynamic Programming Table

Subset of
Interesting Best plan
tables in
-order
FROM clause columns

Cost

{R, S}

<none>

hashjoin(R 1000
,S)

{R, S}

<R.a, S.b>

sortmerge
(R,S)

1500

A Note on Interesting Orders

An intermediate result has an
interesting order if it is sorted
by any of:
ORDER BY attributes
GROUP BY attributes
Join attributes of yet-to-be-added
(downstream) joins

Enumeration of Plans
(Contd.)
An N-1 way plan is not combined with an
additional relation unless there is a join
condition between them, unless all predicates in
WHERE have been used up.
i.e., avoid Cartesian products if possible.
ORDER BY, GROUP BY, aggregates etc. handled as a
final step, using either an `interestingly ordered
plan or an additonal sort/hash operator.
In spite of pruning plan space, this approach is
still exponential in the # of tables.
Recall that in practice, COST considered is #IOs
+ factor * CPU Inst

Example

Sailors:
Hash, B+ on sid
Reserves:
Clustered B+ tree on
bid
B+ on sid
Boats
B+, Hash on color

Select S.sid, COUNT(*) AS number

FROM Sailors S, Reserves R, Boats B
WHERE S.sid = R.sid AND R.bid = B.bid
AND B.color = red
GROUP BY S.sid

Sid, COUNT(*) AS numbes

GROUPBY sid
sid=sid

bid=bid

Color=red

Sailors

Reserves

Boats

Pass1: Best plan(s) for accessing each

relation
Reserves, Sailors: File Scan
Q: What about Clustered B+ on Reserves.bid???
Boats: B+ tree & Hash on color

Pass 1
Best plan for accessing each
relation regarded as the first
relation in an execution plan
Reserves, Sailors: File Scan
Boats: B+ tree & Hash on color

Pass 2
For each of the plans in pass 1, generate plans
joining another relation as the inner, using all join
methods (and matching inner access methods)

File Scan Reserves (outer) with Boats (inner)

File Scan Reserves (outer) with Sailors (inner)
File Scan Sailors (outer) with Boats (inner)
File Scan Sailors (outer) with Reserves (inner)
Boats hash on color with Sailors (inner)
Boats Btree on color with Sailors (inner)
Boats hash on color with Reserves (inner) (sort-merge)
Boats Btree on color with Reserves (inner) (BNL)
Retain cheapest plan for each pair of relations
Q: are there interesting orders?

Pass 3 and beyond

For each of the plans retained from Pass 2,
taken as the outer, generate plans for the
next join
eg Boats hash on color with Reserves (bid) (inner)
(sortmerge))
inner Sailors (B-tree sid) sort-merge

Then, add the cost for doing the group by

and aggregate:
This is the cost to sort the result by sid, unless it
has already been sorted by a previous operator.
Then, choose the cheapest plan

Points to Remember
Must understand optimization in order to
understand the performance impact of a given
database design (relations, indexes) on a
workload (set of queries).
Two parts to optimizing a query:
Consider a set of alternative plans.
Good to prune search space; e.g., left-deep plans only,
avoid Cartesian products.

Must estimate cost of each plan that is considered.

Output cardinality and cost for each plan node.
Key issues: Statistics, indexes, operator implementations.

Points to Remember
Single-relation queries:
All access paths considered, cheapest is chosen.
Issues: Selections that match index, whether index
key has all needed fields and/or provides tuples in a
desired order.

More Points to Remember

Multiple-relation queries:
All single-relation plans are first enumerated.
Selections/projections considered as early as possible.
Next, for each 1-relation plan, all ways of joining another
relation (as inner) are considered.
Next, for each 2-relation plan that is `retained, all ways
of joining another relation (as inner) are considered, etc.
At each level, for each subset of relations, only best plan
for each interesting order of tuples is `retained.

Summary
Optimization is the reason for the lasting
power of the relational system
But it is primitive in some ways
New areas: Smarter summary statistics
(fancy histograms and sketches), autotuning statistics, adaptive runtime reoptimization (e.g. eddies)