Index Compression
David Kauchak
cs160
Fall 2009
adapted from:
http://www.stanford.edu/class/cs276/handouts/lecture5-indexcompression.ppt
Administrative
Homework 2
Assignment 1
Assignment 2
Pair programming?
RCV1 token normalization
size of
word types (terms)
dictionary
Size
(K)
% cumul
%
Unfiltered
484
No numbers
474
-2
-2
Case folding
392 -17
-19
30 stopwords
391
-0
-19
150 stopwords
391
-0
-19
stemming
322 -17
-33
TDT token normalization
normalization
terms
% change
none
120K
number folding
117K
3%
lowercasing
100K
17%
stemming
95K
25%
stoplist
120K
0%
number & lower & stoplist
97K
20%
all
78K
35%
What normalization technique(s) should we use?
Index parameters vs. what we index
size of
word types (terms)
non-positional
postings
positional postings
dictionary
non-positional index
positional index
Size
(K)
Size (K)
% cumul
%
cumul Size (K)
%
109,971
cumul
%
Unfiltered
484
197,879
No numbers
474
-2
-2
100,680
-8
-8
179,158
-9
-9
Case folding
392 -17
-19
96,969
-3
-12
179,158
-9
30 stopwords
391
-0
-19
83,390 -14
-24
121,858 -31
-38
150 stopwords
391
-0
-19
67,002 -30
-39
94,517 -47
-52
stemming
322 -17
-33
63,812
-42
94,517
-52
-4
Index parameters vs. what we index
size of
word types (terms)
non-positional
postings
positional postings
dictionary
non-positional index
positional index
Size
(K)
Size (K)
% cumul
%
cumul Size (K)
%
109,971
cumul
%
Unfiltered
484
197,879
No numbers
474
-2
-2
100,680
-8
-8
179,158
-9
-9
Case folding
392 -17
-19
96,969
-3
-12
179,158
-9
30 stopwords
391
-0
-19
83,390 -14
-24
121,858 -31
-38
150 stopwords
391
-0
-19
67,002 -30
-39
94,517 -47
-52
stemming
322 -17
-33
63,812
-42
94,517
-52
-4
Index parameters vs. what we index
size of
word types (terms)
non-positional
postings
positional postings
dictionary
non-positional index
positional index
Size
(K)
Size (K)
% cumul
%
cumul Size (K)
%
109,971
cumul
%
Unfiltered
484
197,879
No numbers
474
-2
-2
100,680
-8
-8
179,158
-9
-9
Case folding
392 -17
-19
96,969
-3
-12
179,158
-9
30 stopwords
391
-0
-19
83,390 -14
-24
121,858 -31
-38
150 stopwords
391
-0
-19
67,002 -30
-39
94,517 -47
-52
stemming
322 -17
-33
63,812
-42
94,517
-52
-4
Corpora statistics
statistic
documents
avg. # of tokens
per doc
terms
non-positional
postings
TDT
16K
400
Reuters RCV1
800K
200
100K
?
400K
100M
vocabulary size
How does the vocabulary size grow
with the size of the corpus?
number of documents
log of the vocabulary size
How does the vocabulary size grow
with the size of the corpus?
log of the number of documents
Heaps law
Vocab size = k (tokens)b
M = k Tb
Typical values: 30 k 100 and b 0.5.
Does this explain the plot we saw before?
log M= log k + b log(T)
What does this say about the vocabulary size as we
increase the number of documents?
there are almost always new words to be seen: increasing
the number of documents increases the vocabulary size
to get a linear increase in vocab size, need to add
exponential number of documents
How does the vocabulary size grow
with the size of the corpus?
log of the vocabulary size
log10M = 0.49 log10T +
1.64 is the best least
squares fit.
M = 101.64T0.49
k = 101.64 44
b = 0.49.
log of the number of documents
Discussion
How do token normalization techniques
and similar efforts like spelling
correction interact with Heaps law?
Heaps law and compression
Today, were talking about index compression,
i.e. reducing the memory requirement for storing
the index
What implications does Heaps law have for
compression?
Dictionary sizes will continue to increase
Dictionaries can be very large
word frequency
How does a words frequency relate to
its frequency rank?
words frequency rank
log of the frequency
How does a words frequency relate to
its frequency rank?
log of the frequency rank
Zipfs law
In natural language, there are a few very frequent
terms and very many very rare terms
Zipfs law: The ith most frequent term has frequency
proportional to 1/i
frequencyi c/i
where c is a constant
log(frequencyi) log c log i
Consequences of Zipfs law
If the most frequent term (the) occurs cf1
times, how often do the 2nd and 3rd most frequent
occur?
then the second most frequent term (of) occurs
cf1/2 times
the third most frequent term (and) occurs cf1/3
times
If were counting the number of words in a given
frequency range, lowering the frequency band
linearly results in an exponential increase in the
number of words
Zipfs law and compression
What implications does Zipfs law have for
compression?
Some terms will
occur very frequently
in positional postings
lists
word frequency
Dealing with these
well can drastically
reduce the index size
words frequency rank
Index compression
Compression techniques attempt to decrease the
space required to store an index
What other benefits does compression have?
Keep more stuff in memory (increases speed)
Increase data transfer from disk to memory
[read compressed data and decompress] is faster than
[read uncompressed data]
What does this assume?
Decompression algorithms are fast
True of the decompression algorithms we use
Inverted index
word1
word2
wordn
What do we need to store?
How are we storing it?
Compression in inverted indexes
First, we will consider space for dictionary
Make it small enough to keep in main memory
Then the postings
Reduce disk space needed, decrease time to read
from disk
Large search engines keep a significant part of
postings in memory
Lossless vs. lossy compression
What is the difference between lossy and lossless
compression techniques?
Lossless compression: All information is preserved
Lossy compression: Discard some information, but
attempt to keep information that is relevant
Several of the preprocessing steps can be viewed as lossy
compression: case folding, stop words, stemming, number
elimination.
Prune postings entries that are unlikely to turn up in the top k
list for any query
Where else have you seen lossy and lossless
compresion techniques?
Why compress the dictionary
Must keep in memory
Search begins with the dictionary
Memory footprint competition
Embedded/mobile devices
What is a straightforward way of
storing the dictionary?
What is a straightforward way of
storing the dictionary?
Array of fixed-width entries
~400,000 terms; 28 bytes/term = 11.2 MB.
20 bytes
4 bytes each
Fixed-width terms are wasteful
Any problem with this approach?
Most of the bytes in the Term column are wasted
we allot 20 bytes for 1 letter terms
Written English averages ~4.5 characters/word
And we still cant handle supercalifragilisticexpialidocious
Is this the number to use for estimating the
dictionary size?
Ave. dictionary word in English: ~8 characters
Short words dominate token counts but not type
average
Any ideas?
Store the dictionary as one long string
.systilesyzygeticsyzygialsyzygyszaibelyiteszczecinszomo.
Gets ride of wasted space
If the average word is 8 characters, what is our
savings over the 20 byte representation?
Theoretically, 60%
Any issues?
Dictionary-as-a-String
Store dictionary as a (long) string of characters:
Pointer to next word shows end of current word
.systilesyzygeticsyzygialsyzygyszaibelyiteszczecinszomo.
Total string length =
400K x 8B = 3.2MB
Pointers resolve 3.2M
positions: log23.2M =
22bits = 3bytes
How much memory to store the pointers?
Space for dictionary as a string
Fixed-width
As a string
20 bytes per term = 8 MB
6.4 MB (3.2 for dictionary and 3.2 for pointers)
20% reduction!
Still a long way from 60%. Any way we can store
less pointers?
Blocking
Store pointers to every kth term string
.systilesyzygeticsyzygialsyzygyszaibelyiteszczecinszomo.
What else do we need?
Blocking
Store pointers to every kth term string
Example below: k = 4
Need to store term lengths (1 extra byte)
.7systile9syzygetic8syzygial6syzygy11szaibelyite8szczecin9szomo.
Save 9 bytes
on 3
pointers.
Lose 4 bytes on
term lengths.
Net
Where we used 3 bytes/pointer without blocking
3 x 4 = 12 bytes for k=4 pointers,
now we use 3+4=7 bytes for 4 pointers.
Shaved another ~0.5MB; can save more with larger k.
Why not go with larger k?
Dictionary search without blocking
How would we search for a dictionary entry?
.systilesyzygeticsyzygialsyzygyszaibelyiteszczecinszomo.
Dictionary search without blocking
Binary search
Assuming each
dictionary term is equally
likely in query (not really
so in practice!), average
number of comparisons
=?
(1+22+43+4)/8 ~2.6
Dictionary search with blocking
What about with blocking?
.7systile9syzygetic8syzygial6syzygy11szaibelyite8szczecin9szomo.
Dictionary search with blocking
Binary search down to 4-term block
Then linear search through terms in block.
Blocks of 4 (binary tree), avg. = ?
(1+22+23+24+5)/8 = 3 compares
More improvements
8automata8automate9automatic10automation
Were storing the words in sorted order
Any way that we could further compress this
block?
Front coding
Front-coding:
Sorted words commonly have long common prefix
store differences only
(for last k-1 in a block of k)
8automata8automate9automatic10automation
8automat*a1e2ic3ion
Encodes automat
Extra length
beyond automat
Begins to resemble general string compression
RCV1 dictionary compression
Technique
Fixed width
Size in MB
11.2
String with pointers to every term
7.6
Blocking k = 4
7.1
Blocking + front coding
5.9
Postings compression
The postings file is much larger than the
dictionary, by a factor of at least 10
A posting for our purposes is a docID
Regardless of our postings list data structure, we
need to store all of the docIDs
For Reuters (800,000 documents), we would use
32 bits per docID when using 4-byte integers
Alternatively, we can use log2 800,000 20 bits
per docID
Postings: two conflicting forces
Where is most of the storage going?
Frequent terms will occur in most of the
documents and require a lot of space
A term like the occurs in virtually every doc, so
20 bits/posting is too expensive.
Prefer 0/1 bitmap vector in this case
A term like arachnocentric occurs in maybe one
doc out of a million we would like to store this
posting using log2 1M ~ 20 bits.
Postings file entry
We store the list of docs containing a term in
increasing order of docID.
computer: 33,47,154,159,202
Is there another way we could store this sorted
data?
Store gaps: 33,14,107,5,43
14 = 47-33
107 = 154 47
5 = 159 - 154
Fixed-width
How many bits do we need to encode the gaps?
Does this buy us anything?
Variable length encoding
Aim:
For arachnocentric, we will use ~20 bits/gap
entry
For the, we will use ~1 bit/gap entry
Key challenge: encode every integer (gap) with
as few bits as needed for that integer
1, 5, 5000, 1, 1524723,
for smaller integers, use fewer bits
for larger integers, use more bits
Variable length coding
1, 5, 5000, 1, 1124
1, 101, 1001110001, 1, 10001100101
Fixed width:
000000000100000001011001110001
every 10 bits
Variable width:
11011001110001110001100101
?
Variable Byte (VB) codes
Rather than use 20 bits, i.e. record gaps with the
smallest number of bytes to store the gap
1, 101, 1001110001
00000001, 00000101, 00000010 01110001
1 byte
1 byte
2 bytes
00000001000001010000001001110001
?
VB codes
Reserve the first bit of each byte as the
continuation bit
If the bit is 1, then were at the end of the bytes
for the gap
If the bit is 0, there are more bytes to read
1, 101, 1001110001
100000011000010100000100 11110001
For each byte used, how many bits of the gap are
we storing?
Example
docIDs
824
gaps
VB code
00000110
10111000
829
215406
214577
10000101
00001101
00001100
10110001
Postings stored as the byte concatenation
000001101011100010000101000011010000110010110001
Key property: VB-encoded postings are
uniquely prefix-decodable.
For a small gap (5), VB
uses a whole byte.
Other variable codes
Instead of bytes, we can also use a different
unit of alignment: 32 bits (words), 16 bits, 4
bits (nibbles) etc.
What are the pros/cons of a smaller/larger
unit of alignment?
Larger units waste less space on continuation bits
(1 of 32 vs. 1 of 8)
Smaller unites waste less space on encoding
smaller number, e.g. to encode 1 we waste (6
bits vs. 30 bits)
More codes
100000011000010100000100 11110001
Still seems wasteful
What is the major challenge for these variable
length codes?
We need to know the length of the number!
Idea: Encode the length of the number so that
we know how many bits to read
Gamma codes
Represent a gap as a pair length and offset
offset is G in binary, with the leading bit cut off
13 1101 101
17 10001 0001
50 110010 10010
length is the length of offset
13 (offset 101), it is 3
17 (offset 0001), it is 4
50 (offset 10010), it is 5
Encoding the length
Weve stated what the length is, but not how to
encode it
What is a requirement of our length encoding?
Lengths will have variable length (e.g. 3, 4, 5 bits)
We must be able to decode it without any ambiguity
Any ideas?
Unary code
Encode a number n as n 1s, followed by a 0, to
mark the end of it
5 111110
12 1111111111110
Gamma code examples
number
length
0
1
2
3
4
9
13
24
511
1025
offset
-code
Gamma code examples
number
length
offset
-code
none
10
10,0
10
10,1
110
00
110,00
1110
001
1110,001
13
1110
101
1110,101
24
11110
1000
11110,1000
511
111111110
11111111
111111110,11111111
1025
11111111110
0000000001
11111111110,0000000001
Gamma code properties
Uniquely prefix-decodable, like VB
All gamma codes have an odd number of bits
What is the fewest number of bits we could
expect to express a gap (without any other
knowledge of the other gaps)?
log2 (gap)
How many bits do gamma codes use?
2 log2 (gap) +1 bits
Almost within a factor of 2 of best possible
Gamma seldom used in practice
Machines have word boundaries 8, 16, 32 bits
Compressing and manipulating at individual bitgranularity will slow down query processing
Variable byte alignment is potentially more
efficient
Regardless of efficiency, variable byte is
conceptually simpler at little additional space cost
RCV1 compression
Data structure
dictionary, fixed-width
Size in MB
11.2
dictionary, term pointers into string
7.6
with blocking, k = 4
7.1
with blocking & front coding
5.9
collection (text, xml markup etc)
collection (text)
Term-doc incidence matrix
3,600.0
960.0
40,000.0
postings, uncompressed (32-bit words)
400.0
postings, uncompressed (20 bits)
250.0
postings, variable byte encoded
116.0
postings, -encoded
101.0
Index compression summary
We can now create an index for highly efficient
Boolean retrieval that is very space efficient
Only 4% of the total size of the collection
Only 10-15% of the total size of the text in the
collection
However, weve ignored positional information
Hence, space savings are less for indexes used
in practice
But techniques substantially the same
Resources
IIR 5
F. Scholer, H.E. Williams and J. Zobel. 2002.
Compression of Inverted Indexes For Fast Query
Evaluation. Proc. ACM-SIGIR 2002.
V. N. Anh and A. Moffat. 2005. Inverted Index
Compression Using Word-Aligned Binary Codes.
Information Retrieval 8: 151166.
