Download File Organizations and Indexing - Department of Computer Science

File Organizations and Indexing
Chapter 8
Jianping Fan
Dept of Computer Science
UNC-Charlotte
Database Management Systems, R. Ramakrishnan and J. Gehrke
1
How to organize files in disk will affect the
performance of database!
Buffer (RAM)
Data Page
I/O cost
Database Management Systems, R. Ramakrishnan and J. Gehrke
File Storage Disk
2
How client access the file will also affect the
performance of database!
Search Key
Description Schema
Primary Key
attribute
Candidate Key
Why we do not always use primary key as the search key?
Database Management Systems, R. Ramakrishnan and J. Gehrke
3
Alternative File Organizations
Many alternatives exist, each ideal for some
situation , and not so good in others:
–
–
–
Heap files: Suitable when typical access is a file
scan retrieving all records.
Sorted Files: Best if records must be retrieved in
some order, or only a `range’ of records is needed.
Hashed Files: Good for equality selections.
 File is a collection of buckets. Bucket = primary
page plus zero or more overflow pages.
 Hashing function h: h(r) = bucket in which
record r belongs. h looks at only some of the
fields of r, called the search fields.
Database Management Systems, R. Ramakrishnan and J. Gehrke
4
We can sort the objects with only one attribute!
Heap files
Attributes
Sorted files
Attributes
Search key
name
objects
objects
Database Management Systems, R. Ramakrishnan and J. Gehrke
5
We can hash the objects based on only one attribute!
Hash files
Attributes
Search key
buckets
20% empty
attribute
Hash function
Database Management Systems, R. Ramakrishnan and J. Gehrke
6
Cost Model for Our Analysis
I/O cost
Disk
Different file organization techniques may suit for different operations!
Database Management Systems, R. Ramakrishnan and J. Gehrke
7
Cost of Operations
Heap
File
Sorted
File
Hashed
File
Scan all recs
Equality Search
Range Search
Insert
Delete
 Several assumptions underlie these (rough) estimates!
Database Management Systems, R. Ramakrishnan and J. Gehrke
8
Cost Model for Our Analysis
We include/ignore CPU cost analysis:
–
–
–
–
–
B: The number of data pages
R: Number of records per page
D: (Average) time to read or write disk page
Measuring number of page I/O’s ignores gains of
pre-fetching blocks of pages; thus, even I/O cost is
only approximated.
Average-case analysis; based on several simplistic
assumptions.
C: Data Matching time, H: Hashing time
Database Management Systems, R. Ramakrishnan and J. Gehrke
9
Assumptions in Our Analysis
Single record insert and delete.
 Heap Files:

–
–

Sorted Files:
–
–

Equality selection on key; exactly one match.
Insert always at end of file.
Files compacted after deletions.
Selections on sort field(s).
Hashed Files:
–
No overflow buckets, 80% page occupancy.
Database Management Systems, R. Ramakrishnan and J. Gehrke
10
Cost of Operations
Heap files
Attributes
Sorted files
Attributes
name
objects
objects
If search key is same as the attribute for sorting!
Database Management Systems, R. Ramakrishnan and J. Gehrke
11
Cost of Operations
Operations
Scan
Heap File
B(D+RC)
Equal Search
0.5B(D+RC)
Range Search
B(D+RC)
Insert
Delete
Sorted File
B(D+RC)
DlogB + ClogR
(DlogB+ClogR)+D*N
2D + RC
(DlogB+ClogR)+B(D+RC)
D + 0.5B(D+RC)
If not just put it at the end!
(DlogB+ClogR)+B(D+RC)
Cost for re-organization!
Database Management Systems, R. Ramakrishnan and J. Gehrke
12
Cost of Operations
Hash files
Search Key
Attributes
buckets
h = a*value + b mod n
20% empty
attribute
Hash function
First use the search key to
find the Bucket, and the scan
the data in that page.
Database Management Systems, R. Ramakrishnan and J. Gehrke
13
Cost of Operations
Operations
Scan
Hash File
1.25B(D+RC)
Equal Search
H+D+0.5RC
Range Search
1.25B(D+RC)
Insert
Delete
(20% empty)
(20% empty)
C+D+(H+D+0.5RC)
(H+D+0.5RC)+ (C+D)
Database Management Systems, R. Ramakrishnan and J. Gehrke
14
Cost of Operations
Scan all recs
Heap
File
BD
Equality Search 0.5 BD
Only concern I/O cost!
Sorted
File
BD
Hashed
File
1.25 BD
D log2B
D
Range Search
BD
D (log2B + # of 1.25 BD
pages with
matches)
Search + BD
2D
Insert
2D
Delete
Search + D Search + BD
2D
How to choose file organization?
Database Management Systems, R. Ramakrishnan and J. Gehrke
15
What is the Database?
Database Management
Database indexing is
stored in main memory
with several pages
Buffer pool
a. File management
b. Indexing tree
Database Indexing
I/O cost
Data Storage Management
Database Management Systems, R. Ramakrishnan and J. Gehrke
16
Alternatives for Data Entry k* in Index

Three alternatives:
 Data record with key value k
 <k, rid of data record with search key value k>
 <k, list of rids of data records with search key k>

Choice of alternative for data entries is
orthogonal to the indexing technique used to
locate data entries with a given key value k.
–
–
Examples of indexing techniques: B+ trees, hashbased structures
Typically, index contains auxiliary information that
directs searches to the desired data entries
Database Management Systems, R. Ramakrishnan and J. Gehrke
17
Indexes

An index on a file speeds up selections on the
search key fields for the index.
–
–

Any subset of the fields of a relation can be the
search key for an index on the relation.
Search key is not the same as key (minimal set of
fields that uniquely identify a record in a relation).
An index contains a collection of data entries,
and supports efficient retrieval of all data
entries k* with a given key value k.
Database Management Systems, R. Ramakrishnan and J. Gehrke
18
Indexes
How to build database indexing tree?
Root table
Database Management Systems, R. Ramakrishnan and J. Gehrke
19
Alternatives for Data Entries (Contd.)

Alternative 1:
–
–
–
If this is used, index structure is a file organization
for data records (like Heap files or sorted files).
At most one index on a given collection of data
records can use Alternative 1. (Otherwise, data
records duplicated, leading to redundant storage
and potential inconsistency.)
If data records very large, # of pages containing
data entries is high. Implies size of auxiliary
information in the index is also large, typically.
Database Management Systems, R. Ramakrishnan and J. Gehrke
20
Alternatives for Data Entries (Contd.)

Alternatives 2 and 3:
–
–
–
Data entries typically much smaller than data
records. So, better than Alternative 1 with large
data records, especially if search keys are small.
(Portion of index structure used to direct search is
much smaller than with Alternative 1.)
If more than one index is required on a given file, at
most one index can use Alternative 1; rest must use
Alternatives 2 or 3.
Alternative 3 more compact than Alternative 2, but
leads to variable sized data entries even if search
keys are of fixed length.
Database Management Systems, R. Ramakrishnan and J. Gehrke
21
Index Classification

Primary vs. secondary: If search key contains
primary key, then called primary index.
–

Unique index: Search key contains a candidate key.
Clustered vs. unclustered: If order of data records
is the same as, or `close to’, order of data entries,
then called clustered index.
–
–
–
Alternative 1 implies clustered, but not vice-versa.
A file can be clustered on at most one search key.
Cost of retrieving data records through index varies
greatly based on whether index is clustered or not!
Database Management Systems, R. Ramakrishnan and J. Gehrke
22
Clustered vs. Unclustered Index

Suppose that Alternative (2) is used for data entries,
and that the data records are stored in a Heap file.
–
–
To build clustered index, first sort the Heap file (with
some free space on each page for future inserts).
Overflow pages may be needed for inserts. (Thus, order of
data recs is `close to’, but not identical to, the sort order.)
CLUSTERED
Index entries
direct search for
data entries
Data entries
UNCLUSTERED
Data entries
(Index File)
(Data file)
Records
Database Management Systems, R.Data
Ramakrishnan
and J. Gehrke
Data Records
23
Index Classification (Contd.)

Dense vs. Sparse: If
there is at least one data
entry per search key
value (in some data
record), then dense.
–
–
–
Alternative 1 always
leads to dense index.
Every sparse index is
clustered!
Sparse indexes are
smaller; however, some
useful optimizations are
based on dense indexes.
Ashby, 25, 3000
22
Basu, 33, 4003
Bristow, 30, 2007
25
30
Ashby
33
Cass
Cass, 50, 5004
Smith
Daniels, 22, 6003
Jones, 40, 6003
40
44
Smith, 44, 3000
44
50
Tracy, 44, 5004
Sparse Index
on
Name
Database Management Systems, R. Ramakrishnan and J. Gehrke
Data File
Dense Index
on
Age
24
Index Classification (Contd.)

Composite Search Keys: Search
on a combination of fields.
–
Equality query: Every field
value is equal to a constant
value. E.g. wrt <sal,age> index:

–
Range query: Some field value
is not a constant. E.g.:


age=20 and sal =75
age =20; or age=20 and sal > 10
Data entries in index sorted
by search key to support
range queries.
–
–
Lexicographic order, or
Spatial order.
Examples of composite key
indexes using lexicographic order.
11,80
11
12,10
12
12,20
13,75
<age, sal>
10,12
20,12
75,13
name age sal
bob 12
10
cal 11
80
joe 12
20
sue 13
75
13
<age>
10
Data records
sorted by name
80,11
<sal, age>
Data entries in index
sorted by <sal,age>
Database Management Systems, R. Ramakrishnan and J. Gehrke
12
20
75
80
<sal>
Data entries
sorted by <sal>
25
Summary
Many alternative file organizations exist, each
appropriate in some situation.
 If selection queries are frequent, sorting the
file or building an index is important.

–
–

Hash-based indexes only good for equality search.
Sorted files and tree-based indexes best for range
search; also good for equality search. (Files rarely
kept sorted in practice; B+ tree index is better.)
Index is a collection of data entries plus a way
to quickly find entries with given key values.
Database Management Systems, R. Ramakrishnan and J. Gehrke
26
Summary (Contd.)

Data entries can be actual data records, <key,
rid> pairs, or <key, rid-list> pairs.
–
Choice orthogonal to indexing technique used to
locate data entries with a given key value.
Can have several indexes on a given file of
data records, each with a different search key.
 Indexes can be classified as clustered vs.
unclustered, primary vs. secondary, and
dense vs. sparse. Differences have important
consequences for utility/performance.

Database Management Systems, R. Ramakrishnan and J. Gehrke
27
Homework
Condition for comparison: Database has B
data pages, each data page has R objects, I/O
cost for read/write a data page is D, the time
for processing an object is C. If hash function
is used, the hash time is H. Without
considering re-organization!
 (a) If only I/O cost is included, compare the
performances for heap file, sorted file, hashed
file for file organization.
 (b) If CPU cost also includes, compare the
performances for heap file, sorted file, hashed
file for file organization.

Database Management Systems, R. Ramakrishnan and J. Gehrke
28