Bitmap Indexes

 Bitmap Indexing
 Bitmap Indices
 Bitmap indices are a special type of index designed for
efficient querying on multiple keys
 Records in a relation are assumed to be numbered
sequentially from, say, 0
 Given a number n it must be easy to retrieve record n
• Particularly easy if records are of fixed size
 Applicable on attributes that take on a relatively small
number of distinct values
 E.g. gender, country, state, …
 E.g. income-level (income broken up into a small number
of levels such as 0-9999, 10000-19999, 20000-50000, 50000-
infinity)
 A bitmap is simply an array of bits
 Bitmap Indices (Cont.)
 In its simplest form a bitmap index on an attribute has a bitmap for
each value of the attribute
 Bitmap has as many bits as records
 In a bitmap for value v, the bit for a record is 1 if the record has
the value v for the attribute, and is 0 otherwise
 Bitmap Indices (Cont.)
 Bitmap indices are useful for queries on multiple attributes
 not particularly useful for single attribute queries
 Queries are answered using bitmap operations
 Intersection (and)
 Union (or)
 Complementation (not)
 Each operation takes two bitmaps of the same size and applies
the operation on corresponding bits to get the result bitmap
 E.g. 100110 AND 110011 = 100010
100110 OR 110011 = 110111
NOT 100110 = 011001
 Males with income level L1: 10010 AND 10100 = 10000
• Can then retrieve required tuples.
• Counting number of matching tuples is even faster
 Bitmap Index---Definition

 A collection of bit-vectors of length n

 n means the total number of records
 Each bit of this vector stands for one possible
values that may appear in this field.
 At position i for the vector of value v, if the i th
record have value v , then the value of vector bit i
is “1”. otherwise “0”
Example 1.
How the bitmap index forms
 Sample Relation
R(F,G)
F G
Tuple # 1 30 foo
Tuple # 2 30 bar
Tuple # 3 40 baz
Tuple # 4 50 foo
Tuple # 5 40 bar
Tuple # 6 30 baz
 Example 1.
How the bitmap index forms (continue..)
 Let’s see field “F”
 How many DISTINCT values ?
3 (30,40,50)
so, the total number of tuples in the bitmap index
is 3.
 How many records ?
6
so, the length of bit-vector is 6.
Example 1.
How the bitmap index forms (continue..)
 Distinct value: 3
 Total tuple : 6
 Bitmap index:

value Bit-vector
30 xxxxxx
40 xxxxxx
50 xxxxxx

Distinct values: Length of Bit-Vector is 6

Example 1.
How the bitmap index forms (continue..)
 How to form bit- value Bit-vector
vector ? 30 xxxxxx
 Let’s see Value 30
40 xxxxxx
 Which Tuples has
50 xxxxxx
the value 30 in field
F?
(1, 2, 6)

1
Example 1.
How the bitmap index forms (continue..)
 So, the bit-vector of
value 30 is :
Tuple1 Tuple2 Tuple3 Tuple4 Tuple5 Tuple6

1 1 0 0 0 1

1
 Example 1.
How the bitmap index forms
 Fill the bit-map index of Field “F” as following:

value Bit-vector

30 110001

40 001010
50 000100

1
 Do we need Bitmap-Index?

 How to find the i th record in a bitmap index?

1
 Advantage of Bitmap Index

 Accelerate the search.

 Example 2
 Consider following Relation:
Movie (title, year, length, studioName)
 We need to run the following query:

SELECT title
FROM Movie
WHERE studioName = ‘Disney’ AND
year = 1995;
1
Example 2. (contiue…)

 With following tuples:

studioName year
Disney 1995
MGM 1996
DreamFactory 2000
MGM 1995
DreamFactory 1996
Disney 2000
1
 Example 2. (contiue…)
 Suppose we build Bitmap index in the field of
“studioName” and “year”
Bitmap Index on Field “studioName”
Disney 100001
MGM 010100
DreamFactory 001010
Bitmap Index on Field “studioName”
1995 100100
1996 010000
2000 001001
16 1
 Example 2. (contiue…)
 The query is :
SELECT title
FROM Movie
WHERE studioName = ‘Disney’ AND
year = 1995;
 So we INTERSECT the bitmap index with value of
‘Disney” and ‘1995’

100001
AND 100100
100000
• Answer: tuple #1
1
 Bitmap Indices (Cont.)
 Bitmap indices generally very small compared with relation
size
 E.g. if record is 100 bytes, space for a single bitmap is
1/800 of space used by relation.
• If number of distinct attribute values is 8, bitmap is only 1% of
relation size
 Deletion needs to be handled properly
 Existence bitmap to note if there is a valid record at a
record location
 Needed for complementation
• not(A=v): (NOT bitmap-A-v) AND ExistenceBitmap
 Should keep bitmaps for all values, even null value
 To correctly handle SQL null semantics for NOT(A=v):
• intersect above result with (NOT bitmap-A-Null)
1
 Efficient Implementation of Bitmap Operations
 Bitmaps are packed into words; a single word and (a basic
CPU instruction) computes and of 32 or 64 bits at once
 E.g. 1-million-bit maps can be and-ed with just 31,250
instruction
 Counting number of 1s can be done fast by a trick:
 Use each byte to index into a precomputed array of 256
elements each storing the count of 1s in the binary
representation
• Can use pairs of bytes to speed up further at a higher memory
cost
 Add up the retrieved counts
 Bitmaps can be used instead of Tuple-ID lists at leaf levels of B+-trees,
for values that have a large number of matching records
 Worthwhile if > 1/64 of the records have that value, assuming a
tuple-id is 64 bits
 Above technique merges benefits of bitmap and B+-tree indices

1
 Index Definition in SQL
 Create an index
create index <index-name> on <relation-name>
(<attribute-list>)
E.g.: create index b-index on branch(branch_name)
 Use create unique index to indirectly specify and enforce
the condition that the search key is a candidate key is a
candidate key.
 Not really required if SQL unique integrity constraint is
supported
 To drop an index
drop index <index-name>
 Most database systems allow specification of type of
index, and clustering.
2