Chapter 13: Query Processing

 Overview
 Measures of Query Cost
 Selection Operation
 Sorting
 Join Operation
 Other Operations
 Evaluation of Expressions

13.1
 Basic Steps in Query Processing
1. Parsing and translation
2. Optimization
3. Evaluation

13.2
 Basic Steps in Query Processing
(Cont.)

 Parsing and translation

 translate the query into its internal form. This is then
translated into relational algebra.
 Parser checks syntax, verifies relations
 Evaluation
 The query-execution engine takes a query-evaluation plan,
executes that plan, and returns the answers to the query.

13.3
 Basic Steps in Query Processing :
Optimization
 A relational algebra expression may have many equivalent
expressions
 E.g., balance2500(balance(account)) is equivalent to
balance(balance2500(account))
 Each relational algebra operation can be evaluated using one of
several different algorithms
 Correspondingly, a relational-algebra expression can be evaluated in
many ways.
 Annotated expression specifying detailed evaluation strategy is
called an evaluation-plan.
 E.g., can use an index on balance to find accounts with balance < 2500,
 or can perform complete relation scan and discard accounts with balance
 2500

13.4
 Basic Steps: Optimization (Cont.)
 Query Optimization: Amongst all equivalent evaluation plans
choose the one with lowest cost.
 Cost is estimated using statistical information from the
database catalog
 e.g. number of tuples in each relation, size of tuples, etc.

 In this chapter we study

 How to measure query costs
 Algorithms for evaluating relational algebra operations
 How to combine algorithms for individual operations in order to
evaluate a complete expression
 In Chapter 14
 We study how to optimize queries, that is, how to find an evaluation
plan with lowest estimated cost

13.5
 Measures of Query Cost
 Cost is generally measured as total elapsed time for
answering query
 Many factors contribute to time cost
 disk accesses, CPU, or even network communication
 Typically disk access is the predominant cost, and is also
relatively easy to estimate. Measured by taking into
account
 Number of seeks * average-seek-cost
 Number of blocks read * average-block-read-cost
 Number of blocks written * average-block-write-cost
 Cost to write a block is greater than cost to read a block

– data is read back after being written to ensure that

the write was successful

13.6
 Measures of Query Cost (Cont.)
 For simplicity we just use number of block transfers from disk as the
cost measure
 We ignore the difference in cost between sequential and random I/O for
simplicity
 We also ignore CPU costs for simplicity
 Costs depends on the size of the buffer in main memory
 Having more memory reduces need for disk access
 Amount of real memory available to buffer depends on other concurrent
OS processes, and hard to determine ahead of actual execution
 We often use worst case estimates, assuming only the minimum amount
of memory needed for the operation is available
 Real systems take CPU cost into account, differentiate between
sequential and random I/O, and take buffer size into account
 We do not include cost to writing output to disk in our cost
formulae

13.7
 Selection Operation
 File scan – search algorithms that locate and retrieve records that
fulfill a selection condition.
 Algorithm A1 (linear search). Scan each file block and test all
records to see whether they satisfy the selection condition.
 Cost estimate (number of disk blocks scanned) = br
 br denotes number of blocks containing records from relation r
 If selection is on a key attribute, cost = (br /2)
 stop on finding record
 Linear search can be applied regardless of
 selection condition or
 ordering of records in the file, or
 availability of indices

13.8
 Selection Operation (Cont.)
 A2 (binary search). Applicable if selection is an equality
comparison on the attribute on which file is ordered.
 Assume that the blocks of a relation are stored contiguously
 Cost estimate (number of disk blocks to be scanned):
 log2(br) — cost of locating the first tuple by a binary search on
the blocks
 Plus number of blocks containing records that satisfy
selection condition
– Will see how to estimate this cost in Chapter 14

13.9
 Selections Using Indices
 Index scan – search algorithms that use an index
 selection condition must be on search-key of index.
 A3 (primary index on candidate key, equality). Retrieve a single record
that satisfies the corresponding equality condition
 Cost = HTi + 1
 A4 (primary index on nonkey, equality) Retrieve multiple records.
 Records will be on consecutive blocks
 Cost = HTi + number of blocks containing retrieved records
 A5 (equality on search-key of secondary index).
 Retrieve a single record if the search-key is a candidate key
 Cost = HTi + 1
 Retrieve multiple records if search-key is not a candidate key
 Cost = HTi + number of records retrieved
– Can be very expensive!
 each record may be on a different block

– one block access for each retrieved record

13.10
 Selections Involving Comparisons
 Can implement selections of the form AV (r) or A  V(r) by using
 a linear file scan or binary search,
 or by using indices in the following ways:
 A6 (primary index, comparison). (Relation is sorted on A)
 For A  V(r) use index to find first tuple  v and scan relation sequentially
from there
 For AV (r) just scan relation sequentially till first tuple > v; do not use index
 A7 (secondary index, comparison ).
 For A  V(r) use index to find first index entry  v and scan index
sequentially from there, to find pointers to records.
 For AV (r) just scan leaf pages of index finding pointers to records, till first
entry > v
 In either case, retrieve records that are pointed to
– requires an I/O for each record
– Linear file scan may be cheaper if many records are
to be fetched!

13.11
 Implementation of Complex Selections
 Conjunction:    . . .  (r)
1 2 n

 A8 (conjunctive selection using one index).

 Select a combination of i and algorithms A1 through A7 that
results in the least cost fori (r).
 Test other conditions on tuple after fetching it into memory buffer.

 A9 (conjunctive selection using multiple-key index).

 Use appropriate composite (multiple-key) index if available.
 A10 (conjunctive selection by intersection of identifiers).
 Requires indices with record pointers.
 Use corresponding index for each condition, and take intersection
of all the obtained sets of record pointers.
 Then fetch records from file
 If some conditions do not have appropriate indices, apply test in
memory.

13.12
 Algorithms for Complex Selections
 Disjunction: 1 2 . . . n (r).
 
 A11 (disjunctive selection by union of identifiers).
 Applicable if all conditions have available indices.
 Otherwise use linear scan.
 Use corresponding index for each condition, and take union of all the
obtained sets of record pointers.
 Then fetch records from file
 Negation: (r)
 Use linear scan on file
 If very few records satisfy , and an index is applicable to 
 Find satisfying records using index and fetch from file

13.13
 Sorting
 We may build an index on the relation, and then use the index to
read the relation in sorted order. May lead to one disk block
access for each tuple.
 For relations that fit in memory, techniques like quicksort can be
used. For relations that don’t fit in memory, external
sort-merge is a good choice.

13.14
 External Sort-Merge
Let M denote memory size (in pages).
1. Create sorted runs. Let i be 0 initially.
Repeatedly do the following till the end of the relation:
(a) Read M blocks of relation into memory
(b) Sort the in-memory blocks
(c) Write sorted data to run Ri; increment i.
Let the final value of I be N
2. Merge the runs (N-way merge). We assume (for now) that N < M.
1. Use N blocks of memory to buffer input runs, and 1 block to buffer
output. Read the first block of each run into its buffer page
2. repeat
1. Select the first record (in sort order) among all buffer pages
2. Write the record to the output buffer. If the output buffer is full
write it to disk.
3. Delete the record from its input buffer page.
If the buffer page becomes empty then
read the next block (if any) of the run into the buffer.
3. until all input buffer pages are empty:

13.15
 External Sort-Merge (Cont.)
 If i  M, several merge passes are required.
 In each pass, contiguous groups of M - 1 runs are
merged.
 A pass reduces the number of runs by a factor of M -1,
and creates runs longer by the same factor.
 E.g. If M=11, and there are 90 runs, one pass
reduces the number of runs to 9, each 10 times the
size of the initial runs
 Repeated passes are performed till all runs have been
merged into one.

13.16
Example: External Sorting Using Sort-Merge

13.17
 External Merge Sort (Cont.)
 Cost analysis:
 Total number of merge passes required: logM–1(br/M).
 Disk accesses for initial run creation as well as in each pass is 2br
 for final pass, we don’t count write cost

– we ignore final write cost for all operations since the output of
an operation may be sent to the parent operation without
being written to disk
Thus total number of disk accesses for external sorting:
br ( 2 logM–1(br / M) + 1)

13.18
 Join Operation
 Several different algorithms to implement joins
 Nested-loop join
 Block nested-loop join
 Indexed nested-loop join
 Merge-join
 Hash-join
 Choice based on cost estimate
 Examples use the following information
 Number of records of customer: 10,000 depositor: 5000
 Number of blocks of customer: 400 depositor: 100

13.19
 Nested-Loop Join
 To compute the theta join r  s
for each tuple tr in r do begin
for each tuple ts in s do begin
test pair (tr,ts) to see if they satisfy the join condition 
if they do, add tr • ts to the result.
end
end
 r is called the outer relation and s the inner relation of the join.
 Requires no indices and can be used with any kind of join
condition.
 Expensive since it examines every pair of tuples in the two
relations.

13.20
 Nested-Loop Join (Cont.)
 In the worst case, if there is enough memory only to hold
one block of each relation, the estimated cost is
nr  bs + br
disk accesses.
 If the smaller relation fits entirely in memory, use that as the
inner relation. Reduces cost to br + bs disk accesses.
 Assuming worst case memory availability cost estimate is
 5000  400 + 100 = 2,000,100 disk accesses with depositor as
outer relation, and
 1000  100 + 400 = 1,000,400 disk accesses with customer as
the outer relation.
 If smaller relation (depositor) fits entirely in memory, the
cost estimate will be 500 disk accesses.
 Block nested-loops algorithm (next slide) is preferable.

13.21
 Block Nested-Loop Join
 Variant of nested-loop join in which every block of inner
relation is paired with every block of outer relation.
for each block Br of r do begin
for each block Bs of s do begin
for each tuple tr in Br do begin
for each tuple ts in Bs do begin
Check if (tr,ts) satisfy the join condition
if they do, add tr • ts to the result.
end
end
end
end

13.22
 Block Nested-Loop Join (Cont.)
 Worst case estimate: br  bs + br block accesses.
 Each block in the inner relation s is read once for each block in the
outer relation (instead of once for each tuple in the outer relation
 Best case: br + bs block accesses.
 Improvements to nested loop and block nested loop
algorithms:
 In block nested-loop, use M — 2 disk blocks as blocking unit for
outer relations, where M = memory size in blocks; use remaining
two blocks to buffer inner relation and output
 Cost = br / (M-2)  bs + br
 If equi-join attribute forms a key or inner relation, stop inner loop
on first match
 Scan inner loop forward and backward alternately, to make use of
the blocks remaining in buffer (with LRU replacement)
 Use index on inner relation if available (next slide)

13.23
 Indexed Nested-Loop Join
 Index lookups can replace file scans if
 join is an equi-join or natural join and
 an index is available on the inner relation’s join attribute
 Can construct an index just to compute a join.
 For each tuple tr in the outer relation r, use the index to look up
tuples in s that satisfy the join condition with tuple tr.
 Worst case: buffer has space for only one page of r, and, for each
tuple in r, we perform an index lookup on s.
 Cost of the join: br + nr  c
 Where c is the cost of traversing index and fetching all matching s
tuples for one tuple or r
 c can be estimated as cost of a single selection on s using the join
condition.
 If indices are available on join attributes of both r and s,
use the relation with fewer tuples as the outer relation.

13.24
 Example of Nested-Loop Join Costs
 Compute depositor customer, with depositor as the outer
relation.
 Let customer have a primary B+-tree index on the join attribute
customer-name, which contains 20 entries in each index node.
 Since customer has 10,000 tuples, the height of the tree is 4, and
one more access is needed to find the actual data
 depositor has 5000 tuples
 Cost of block nested loops join
 400*100 + 100 = 40,100 disk accesses assuming worst case
memory (may be significantly less with more memory)
 Cost of indexed nested loops join
 100 + 5000 * 5 = 25,100 disk accesses.
 CPU cost likely to be less than that for block nested loops join

13.25
 Merge-Join
1. Sort both relations on their join attribute (if not already sorted on the
join attributes).
2. Merge the sorted relations to join them
1. Join step is similar to the merge stage of the sort-merge algorithm.
2. Main difference is handling of duplicate values in join attribute — every
pair with same value on join attribute must be matched

13.26
 Merge-Join (Cont.)
 Can be used only for equi-joins and natural joins
 Each block needs to be read only once (assuming all tuples for
any given value of the join attributes fit in memory
 Thus number of block accesses for merge-join is
b r + bs + the cost of sorting if relations are unsorted.
 hybrid merge-join: If one relation is sorted, and the other has a
secondary B+-tree index on the join attribute
 Merge the sorted relation with the leaf entries of the B+-tree .
 Sort the result on the addresses of the unsorted relation’s tuples
 Scan the unsorted relation in physical address order and merge with
previous result, to replace addresses by the actual tuples
 Sequential scan more efficient than random lookup

13.27
 Hash-Join
 Applicable for equi-joins and natural joins.
 A hash function h is used to partition tuples of both relations
 h maps JoinAttrs values to {0, 1, ..., n}, where JoinAttrs denotes
the common attributes of r and s used in the natural join.
 r0, r1, . . ., rn denote partitions of r tuples
 Each tuple tr  r is put in partition ri where i = h(tr [JoinAttrs]).
 r0,, r1. . ., rn denotes partitions of s tuples
 Each tuple ts s is put in partition si, where i = h(ts [JoinAttrs]).

 Note: In book, ri is denoted as Hri, si is denoted as Hsi and

n is denoted as nh.

13.28
Hash-Join (Cont.)

13.29
 Hash-Join (Cont.)

 r tuples in ri need only to be compared with s tuples in si

Need not be compared with s tuples in any other partition,
since:
 an r tuple and an s tuple that satisfy the join condition will have
the same value for the join attributes.
 If that value is hashed to some value i, the r tuple has to be in ri
and the s tuple in si.

13.30
 Hash-Join Algorithm
The hash-join of r and s is computed as follows.
1. Partition the relation s using hashing function h. When
partitioning a relation, one block of memory is reserved as
the output buffer for each partition.
2. Partition r similarly.
3. For each i:
(a)Load si into memory and build an in-memory hash index on it
using the join attribute. This hash index uses a different hash
function than the earlier one h.
(b)Read the tuples in ri from the disk one by one. For each tuple tr
locate each matching tuple ts in si using the in-memory hash
index. Output the concatenation of their attributes.
Relation s is called the build input and
r is called the probe input.

13.31
 Hash-Join algorithm (Cont.)
 The value n and the hash function h is chosen such that each
si should fit in memory.
 Typically n is chosen as bs/M * f where f is a “fudge factor”,
typically around 1.2
 The probe relation partitions si need not fit in memory
 Recursive partitioning required if number of partitions n is
greater than number of pages M of memory.
 instead of partitioning n ways, use M – 1 partitions for s
 Further partition the M – 1 partitions using a different hash
function
 Use same partitioning method on r
 Rarely required: e.g., recursive partitioning not needed for
relations of 1GB or less with memory size of 2MB, with block size
of 4KB.

13.32
 Handling of Overflows
 Hash-table overflow occurs in partition si if si does not fit in
memory. Reasons could be
 Many tuples in s with same value for join attributes
 Bad hash function
 Partitioning is said to be skewed if some partitions have
significantly more tuples than some others
 Overflow resolution can be done in build phase
 Partition si is further partitioned using different hash function.
 Partition ri must be similarly partitioned.
 Overflow avoidance performs partitioning carefully to avoid
overflows during build phase
 E.g. partition build relation into many partitions, then combine them
 Both approaches fail with large numbers of duplicates
 Fallback option: use block nested loops join on overflowed
partitions

13.33
 Cost of Hash-Join
 If recursive partitioning is not required: cost of hash join is
3(br + bs) +2  nh
 If recursive partitioning required, number of passes required for
partitioning s is logM–1(bs) – 1. This is because each final
partition of s should fit in memory.
 The number of partitions of probe relation r is the same as that
for build relation s; the number of passes for partitioning of r is
also the same as for s.
 Therefore it is best to choose the smaller relation as the build
relation.
 Total cost estimate is:
2(br + bs logM–1(bs) – 1 + br + bs
 If the entire build input can be kept in main memory, n can be set
to 0 and the algorithm does not partition the relations into
temporary files. Cost estimate goes down to br + bs.

13.34
 Example of Cost of Hash-Join
customer depositor
 Assume that memory size is 20 blocks
 bdepositor= 100 and bcustomer = 400.
 depositor is to be used as build input. Partition it into five partitions,
each of size 20 blocks. This partitioning can be done in one pass.
 Similarly, partition customer into five partitions,each of size 80.
This is also done in one pass.
 Therefore total cost: 3(100 + 400) = 1500 block transfers
 ignores cost of writing partially filled blocks

13.35
 Hybrid Hash–Join
 Useful when memory sized are relatively large, and the build input
is bigger than memory.
 Main feature of hybrid hash join:
Keep the first partition of the build relation in memory.
 E.g. With memory size of 25 blocks, depositor can be partitioned
into five partitions, each of size 20 blocks.
 Division of memory:
 The first partition occupies 20 blocks of memory
 1 block is used for input, and 1 block each for buffering the other 4
partitions.
 customer is similarly partitioned into five partitions each of size 80;
the first is used right away for probing, instead of being written out
and read back.
 Cost of 3(80 + 320) + 20 +80 = 1300 block transfers for
hybrid hash join, instead of 1500 with plain hash-join.
 Hybrid hash-join most useful if M >> bs

13.36
 Complex Joins
 Join with a conjunctive condition:

r 1  2...   n s
 Either use nested loops/block nested loops, or
 Compute the result of one of the simpler joins r i s
 final result comprises those tuples in the intermediate result
that satisfy the remaining conditions
1  . . .  i –1  i +1  . . .  n
 Join with a disjunctive condition

r 1  2 ...  n s
 Either use nested loops/block nested loops, or
 Compute as the union of the records in individual joins r  i s:

(r 1 s)  (r 2 s)  . . .  (r n s)

13.37
 Other Operations
 Duplicate elimination can be implemented via hashing
or sorting.
 On sorting duplicates will come adjacent to each other,
and all but one set of duplicates can be deleted.
Optimization: duplicates can be deleted during run
generation as well as at intermediate merge steps in
external sort-merge.
 Hashing is similar – duplicates will come into the same
bucket.
 Projection is implemented by performing projection on
each tuple followed by duplicate elimination.

13.38
 Other Operations : Aggregation
 Aggregation can be implemented in a manner similar to duplicate
elimination.
 Sorting or hashing can be used to bring tuples in the same group
together, and then the aggregate functions can be applied on each
group.
 Optimization: combine tuples in the same group during run
generation and intermediate merges, by computing partial aggregate
values
 For count, min, max, sum: keep aggregate values on tuples
found so far in the group.
– When combining partial aggregate for count, add up the
aggregates
 For avg, keep sum and count, and divide sum by count at the
end

13.39
 Other Operations : Set Operations
 Set operations (,  and ): can either use variant of merge-
join after sorting, or variant of hash-join.
 E.g., Set operations using hashing:

1. Partition both relations using the same hash function, thereby

creating, r1, .., rn r0, and s1, s2.., sn
2. Process each partition i as follows. Using a different hashing
function, build an in-memory hash index on ri after it is brought
into memory.
3. – r  s: Add tuples in si to the hash index if they are not already in
it. At end of si add the tuples in the hash index to the result.
– r  s: output tuples in si to the result if they are already there in
the hash index.
– r – s: for each tuple in si, if it is there in the hash index, delete it
from the index. At end of si add remaining tuples in the hash
index to the result.

13.40
 Other Operations : Outer Join
 Outer join can be computed either as
 A join followed by addition of null-padded non-participating tuples.
 by modifying the join algorithms.
 Modifying merge join to compute r s
 In r s, non participating tuples are those in r – R(r s)
 Modify merge-join to compute r s: During merging, for every
tuple tr from r that do not match any tuple in s, output tr padded with
nulls.
 Right outer-join and full outer-join can be computed similarly.
 Modifying hash join to compute r s
 If r is probe relation, output non-matching r tuples padded with nulls
 If r is build relation, when probing keep track of which
r tuples matched s tuples. At end of si output
non-matched r tuples padded with nulls

13.41
 Evaluation of Expressions
 So far: we have seen algorithms for individual operations
 Alternatives for evaluating an entire expression tree
 Materialization: generate results of an expression whose inputs are
relations or are already computed, materialize (store) it on disk.
Repeat.
 Pipelining: pass on tuples to parent operations even as an operation
is being executed
 We study above alternatives in more detail

13.42
 Materialization
 Materialized evaluation: evaluate one operation at a
time, starting at the lowest-level. Use intermediate
results materialized into temporary relations to evaluate
next-level operations.
 E.g., in figure below, compute and store
 balance 2500 (account )
then compute the store its join with customer, and finally
compute the projections on customer-name.

13.43
 Materialization (Cont.)
 Materialized evaluation is always applicable
 Cost of writing results to disk and reading them back can be
quite high
 Our cost formulas for operations ignore cost of writing results to disk,
so
 Overall cost = Sum of costs of individual operations +
cost of writing intermediate results to disk
 Double buffering: use two output buffers for each operation,
when one is full write it to disk while the other is getting filled
 Allows overlap of disk writes with computation and reduces
execution time

13.44
 Pipelining
 Pipelined evaluation : evaluate several operations simultaneously,
passing the results of one operation on to the next.
 E.g., in previous expression tree, don’t store result of

 balance 2500 (account )

 instead, pass tuples directly to the join.. Similarly, don’t store result of
join, pass tuples directly to projection.
 Much cheaper than materialization: no need to store a temporary
relation to disk.
 Pipelining may not always be possible – e.g., sort, hash-join.
 For pipelining to be effective, use evaluation algorithms that
generate output tuples even as tuples are received for inputs to the
operation.
 Pipelines can be executed in two ways: demand driven and
producer driven

13.45
 Pipelining (Cont.)
 In demand driven or lazy evaluation
 system repeatedly requests next tuple from top level operation
 Each operation requests next tuple from children operations as required, in order to
output its next tuple
 In between calls, operation has to maintain “state” so it knows what to return next
 Each operation is implemented as an iterator implementing the following operations
 open()

– E.g. file scan: initialize file scan, store pointer to beginning of file as state
– E.g.merge join: sort relations and store pointers to beginning of sorted
relations as state
 next()
– E.g. for file scan: Output next tuple, and advance and store file pointer
– E.g. for merge join: continue with merge from earlier state till
next output tuple is found. Save pointers as iterator state.
 close()

13.46
 Pipelining (Cont.)
 In produce-driven or eager pipelining
 Operators produce tuples eagerly and pass them up to their parents
 Buffer maintained between operators, child puts tuples in buffer,
parent removes tuples from buffer
 if buffer is full, child waits till there is space in the buffer, and then
generates more tuples
 System schedules operations that have space in output buffer and
can process more input tuples

13.47
 Evaluation Algorithms for Pipelining
 Some algorithms are not able to output results even as they get
input tuples
 E.g. merge join, or hash join
 These result in intermediate results being written to disk and then read
back always
 Algorithm variants are possible to generate (at least some) results
on the fly, as input tuples are read in
 E.g. hybrid hash join generates output tuples even as probe relation
tuples in the in-memory partition (partition 0) are read in
 Pipelined join technique: Hybrid hash join, modified to buffer partition
0 tuples of both relations in-memory, reading them as they become
available, and output results of any matches between partition 0 tuples
 When a new r0 tuple is found, match it with existing s0 tuples,
output matches, and save it in r0
 Symmetrically for s tuples
0

13.48
 Complex Joins
 Join involving three relations: loan depositor customer
 Strategy 1. Compute depositor customer; use result to compute
loan (depositor customer)
 Strategy 2. Compute loan depositor first, and then join the result
with customer.
 Strategy 3. Perform the pair of joins at once. Build and index on
loan for loan-number, and on customer for customer-name.
 For each tuple t in depositor, look up the corresponding tuples in
customer and the corresponding tuples in loan.
 Each tuple of deposit is examined exactly once.
 Strategy 3 combines two operations into one special-purpose
operation that is more efficient than implementing two joins of two
relations.

13.49