External Sorting
Module 2, Lecture 6
Database Management Systems, R. Ramakrishnan
Why Sort?
A classic problem in computer science!
Data requested in sorted order
e.g., find students in increasing gpa order
Sorting is first step in bulk loading B+ tree index.
Sorting useful for eliminating duplicate copies in a
collection of records (Why?)
Sort-merge join algorithm involves sorting.
Problem: sort 1Gb of data with 1Mb of RAM.
why not virtual memory?
Database Management Systems, R. Ramakrishnan
2-Way Sort: Requires 3 Buffers
Pass 1: Read a page, sort it, write it.
only one buffer page is used
Pass 2, 3, , etc.:
three buffer pages used.
INPUT 1
OUTPUT
INPUT 2
Disk
Main memory buffers
Database Management Systems, R. Ramakrishnan
Disk
3
Two-Way External Merge Sort
Each pass we read + write
each page in file.
N pages in the file => the
number of passes
= log2 N + 1
So toal cost is:
6,2
9,4
8,7
5,6
3,1
3,4
2,6
4,9
7,8
5,6
1,3
4,7
2,3
4,6
1,3
5,6
8,9
Input file
PASS 0
1-page runs
PASS 1
2
2-page runs
PASS 2
2,3
2 N log 2 N + 1
3,4
Idea: Divide and conquer:
sort subfiles and merge
Database Management Systems, R. Ramakrishnan
4,4
6,7
1,2
3,5
6
8,9
4-page runs
PASS 3
1,2
2,3
3,4
8-page runs
4,5
6,6
7,8
9
General External Merge Sort
More than 3 buffer pages. How can we utilize them?
To sort a file with N pages using B buffer pages:
Pass 0: use B buffer pages. Produce N / B sorted runs of B
pages each.
Pass 2, , etc.: merge B-1 runs.
INPUT 1
...
INPUT 2
...
OUTPUT
...
INPUT B-1
Disk
B Main memory buffers
Database Management Systems, R. Ramakrishnan
Disk
5
Cost of External Merge Sort
Number of passes: 1 + log B 1 N / B
Cost = 2N * (# of passes)
E.g., with 5 buffer pages, to sort 108 page file:
Pass 0: 108 / 5 = 22 sorted runs of 5 pages each
(last run is only 3 pages)
Pass 1: 22 / 4 = 6 sorted runs of 20 pages each
(last run is only 8 pages)
Pass 2: 2 sorted runs, 80 pages and 28 pages
Pass 3: Sorted file of 108 pages
Database Management Systems, R. Ramakrishnan
Number of Passes of External Sort
N
B=3 B=5 B=9
100
7
4
3
1,000
10
5
4
10,000
13
7
5
100,000
17
9
6
1,000,000
20
10
7
10,000,000
23
12
8
100,000,000
26
14
9
1,000,000,000 30
15
10
Database Management Systems, R. Ramakrishnan
B=17 B=129 B=257
2
1
1
3
2
2
4
2
2
5
3
3
5
3
3
6
4
3
7
4
4
8
5
4
7
Internal Sort Algorithm
Quicksort is a fast way to sort in memory.
An alternative is tournament sort (a.k.a.
heapsort)
Top: Read in B blocks
Output: move smallest record to output buffer
Read in a new record r
insert r into heap
if r not smallest, then GOTO Output
else remove r from heap
output heap in order; GOTO Top
Database Management Systems, R. Ramakrishnan
More on Heapsort
Fact: average length of a run in heapsort is 2B
The snowplow analogy
Worst-Case:
What is min length of a run?
How does this arise?
Best-Case:
What is max length of a run?
How does this arise?
Quicksort is faster, but ...
Database Management Systems, R. Ramakrishnan
I/O for External Merge Sort
longer runs often means fewer passes!
Actually, do I/O a page at a time
In fact, read a block of pages sequentially!
Suggests we should make each buffer
(input/output) be a block of pages.
But this will reduce fan-out during merge passes!
In practice, most files still sorted in 2-3 passes.
Database Management Systems, R. Ramakrishnan
10
Number of Passes of Optimized Sort
N
100
1,000
10,000
100,000
1,000,000
10,000,000
100,000,000
1,000,000,000
B=1,000
1
1
2
3
3
4
5
5
B=5,000
1
1
2
2
2
3
3
4
B=10,000
1
1
1
2
2
3
3
3
Block size = 32, initial pass produces runs of size 2B.
Database Management Systems, R. Ramakrishnan
11
Double Buffering
To reduce wait time for I/O request to
complete, can prefetch into `shadow block.
Potentially, more passes; in practice, most files still
sorted in 2-3 passes.
INPUT 1
INPUT 1'
INPUT 2
INPUT 2'
OUTPUT
OUTPUT'
Disk
INPUT k
block size
Disk
INPUT k'
B main memory buffers, k-way merge
Database Management Systems, R. Ramakrishnan
12
Sorting Records!
Sorting has become a blood sport!
Parallel sorting is the name of the game ...
Datamation: Sort 1M records of size 100 bytes
Typical DBMS: 15 minutes
World record: 3.5 seconds
12-CPU SGI machine, 96 disks, 2GB of RAM
New benchmarks proposed:
Minute Sort: How many can you sort in 1 minute?
Dollar Sort: How many can you sort for $1.00?
Database Management Systems, R. Ramakrishnan
13
Using B+ Trees for Sorting
Scenario: Table to be sorted has B+ tree index on
sorting column(s).
Idea: Can retrieve records in order by traversing
leaf pages.
Is this a good idea?
Cases to consider:
B+ tree is clustered
B+ tree is not clustered
Database Management Systems, R. Ramakrishnan
Good idea!
Could be a very bad idea!
14
Clustered B+ Tree Used for Sorting
Cost: root to the leftmost leaf, then retrieve
all leaf pages
(Alternative 1)
If Alternative 2 is used?
Additional cost of
retrieving data records:
each page fetched just
once.
Index
(Directs search)
AAAAAAAAAAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAA
AAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAAAAAAAAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAA
AAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAA
AAAAAAAAAAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAA
AAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAA
Data Entries
("Sequence set")
Data Records
Always better than external sorting!
Database Management Systems, R. Ramakrishnan
15
Unclustered B+ Tree Used for Sorting
Alternative (2) for data entries; each data
entry contains rid of a data record. In general,
one I/O per data record!
Index
(Directs search)
AAAAAAAA
AAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAA
AAAAAAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAAAAAA
AAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAA
AAAA
AAAAAAAA
AAAAAAAA
Data Entries
("Sequence set")
Data Records
Database Management Systems, R. Ramakrishnan
16
External Sorting vs. Unclustered Index
N
100
1,000
10,000
100,000
1,000,000
10,000,000
Sorting
p=1
p=10
p=100
200
2,000
40,000
600,000
8,000,000
80,000,000
100
1,000
10,000
100,000
1,000,000
10,000,000
1,000
10,000
100,000
1,000,000
10,000,000
100,000,000
10,000
100,000
1,000,000
10,000,000
100,000,000
1,000,000,000
p: # of records per page
B=1,000 and block size=32 for sorting
p=100 is the more realistic value.
Database Management Systems, R. Ramakrishnan
17
Summary
External sorting is important; DBMS may dedicate
part of buffer pool for sorting!
External merge sort minimizes disk I/O cost:
Pass 0: Produces sorted runs of size B (# buffer pages).
Later passes: merge runs.
# of runs merged at a time depends on B, and block size.
Larger block size means less I/O cost per page.
Larger block size means smaller # runs merged.
In practice, # of runs rarely more than 2 or 3.
Database Management Systems, R. Ramakrishnan
18
Summary, cont.
Choice of internal sort algorithm may matter:
Quicksort: Quick!
Heap/tournament sort: slower (2x), longer runs
The best sorts are wildly fast:
Despite 40+ years of research, were still
improving!
Clustered B+ tree is good for sorting;
unclustered tree is usually very bad.
Database Management Systems, R. Ramakrishnan
19
