<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<meta name="generator" content=
"HTML Tidy for Linux/x86 (vers 12 April 2005), see www.w3.org" />
<title>Hash-Based Containers</title>
<meta http-equiv="Content-Type" content=
"text/html; charset=us-ascii" />
</head>
<body>
<div id="page">
<h1>Hash Table Design</h1>
<h2><a name="overview" id="overview">Overview</a></h2>
<p>The collision-chaining hash-based container has the
following declaration.</p>
<pre>
<b>template</b><
<b>typename</b> Key,
<b>typename</b> Mapped,
<b>typename</b> Hash_Fn = std::hash<Key>,
<b>typename</b> Eq_Fn = std::equal_to<Key>,
<b>typename</b> Comb_Hash_Fn = <a href=
"direct_mask_range_hashing.html">direct_mask_range_hashing</a><>
<b>typename</b> Resize_Policy = <i>default explained below.</i>
<b>bool</b> Store_Hash = <b>false</b>,
<b>typename</b> Allocator = std::allocator<<b>char</b>> >
<b>class</b> <a href=
"cc_hash_table.html">cc_hash_table</a>;
</pre>
<p>The parameters have the following meaning:</p>
<ol>
<li><tt>Key</tt> is the key type.</li>
<li><tt>Mapped</tt> is the mapped-policy, and is explained in
<a href="tutorial.html#assoc_ms">Tutorial::Associative
Containers::Associative Containers Others than Maps</a>.</li>
<li><tt>Hash_Fn</tt> is a key hashing functor.</li>
<li><tt>Eq_Fn</tt> is a key equivalence functor.</li>
<li><tt>Comb_Hash_Fn</tt> is a <i>range-hashing_functor</i>;
it describes how to translate hash values into positions
within the table. This is described in <a href=
"#hash_policies">Hash Policies</a>.</li>
<li><tt>Resize_Policy</tt> describes how a container object
should change its internal size. This is described in
<a href="#resize_policies">Resize Policies</a>.</li>
<li><tt>Store_Hash</tt> indicates whether the hash value
should be stored with each entry. This is described in
<a href="#policy_interaction">Policy Interaction</a>.</li>
<li><tt>Allocator</tt> is an allocator
type.</li>
</ol>
<p>The probing hash-based container has the following
declaration.</p>
<pre>
<b>template</b><
<b>typename</b> Key,
<b>typename</b> Mapped,
<b>typename</b> Hash_Fn = std::hash<Key>,
<b>typename</b> Eq_Fn = std::equal_to<Key>,
<b>typename</b> Comb_Probe_Fn = <a href=
"direct_mask_range_hashing.html">direct_mask_range_hashing</a><>
<b>typename</b> Probe_Fn = <i>default explained below.</i>
<b>typename</b> Resize_Policy = <i>default explained below.</i>
<b>bool</b> Store_Hash = <b>false</b>,
<b>typename</b> Allocator = std::allocator<<b>char</b>> >
<b>class</b> <a href=
"gp_hash_table.html">gp_hash_table</a>;
</pre>
<p>The parameters are identical to those of the
collision-chaining container, except for the following.</p>
<ol>
<li><tt>Comb_Probe_Fn</tt> describes how to transform a probe
sequence into a sequence of positions within the table.</li>
<li><tt>Probe_Fn</tt> describes a probe sequence policy.</li>
</ol>
<p>Some of the default template values depend on the values of
other parameters, and are explained in <a href=
"#policy_interaction">Policy Interaction</a>.</p>
<h2><a name="hash_policies" id="hash_policies">Hash
Policies</a></h2>
<h3><a name="general_terms" id="general_terms">General
Terms</a></h3>
<p>Following is an explanation of some functions which hashing
involves. Figure <a href=
"#hash_ranged_hash_range_hashing_fns">Hash functions,
ranged-hash functions, and range-hashing functions</a>)
illustrates the discussion.</p>
<h6 class="c1"><a name="hash_ranged_hash_range_hashing_fns" id=
"hash_ranged_hash_range_hashing_fns"><img src=
"hash_ranged_hash_range_hashing_fns.png" alt=
"no image" /></a></h6>
<h6 class="c1">Hash functions, ranged-hash functions, and
range-hashing functions.</h6>
<p>Let <i>U</i> be a domain (<i>e.g.</i>, the integers, or the
strings of 3 characters). A hash-table algorithm needs to map
elements of <i>U</i> "uniformly" into the range <i>[0,..., m -
1]</i> (where <i>m</i> is a non-negative integral value, and
is, in general, time varying). <i>I.e.</i>, the algorithm needs
a <i>ranged-hash</i> function</p>
<p><i>f : U × Z<sub>+</sub> → Z<sub>+</sub></i>
,</p>
<p>such that for any <i>u</i> in <i>U</i> ,</p>
<p><i>0 ≤ f(u, m) ≤ m - 1</i> ,</p>
<p>and which has "good uniformity" properties [<a href=
"references.html#knuth98sorting">knuth98sorting</a>]. One
common solution is to use the composition of the hash
function</p>
<p><i>h : U → Z<sub>+</sub></i> ,</p>
<p>which maps elements of <i>U</i> into the non-negative
integrals, and</p>
<p class="c2">g : Z<sub>+</sub> × Z<sub>+</sub> →
Z<sub>+</sub>,</p>
<p>which maps a non-negative hash value, and a non-negative
range upper-bound into a non-negative integral in the range
between 0 (inclusive) and the range upper bound (exclusive),
<i>i.e.</i>, for any <i>r</i> in <i>Z<sub>+</sub></i>,</p>
<p><i>0 ≤ g(r, m) ≤ m - 1</i> .</p>
<p>The resulting ranged-hash function, is</p>
<p><i><a name="ranged_hash_composed_of_hash_and_range_hashing"
id="ranged_hash_composed_of_hash_and_range_hashing">f(u , m) =
g(h(u), m)</a></i> (1) .</p>
<p>From the above, it is obvious that given <i>g</i> and
<i>h</i>, <i>f</i> can always be composed (however the converse
is not true). The STL's hash-based containers allow specifying
a hash function, and use a hard-wired range-hashing function;
the ranged-hash function is implicitly composed.</p>
<p>The above describes the case where a key is to be mapped
into a <i>single position</i> within a hash table, <i>e.g.</i>,
in a collision-chaining table. In other cases, a key is to be
mapped into a <i>sequence of positions</i> within a table,
<i>e.g.</i>, in a probing table. Similar terms apply in this
case: the table requires a <i>ranged probe</i> function,
mapping a key into a sequence of positions withing the table.
This is typically achieved by composing a <i>hash function</i>
mapping the key into a non-negative integral type, a
<i>probe</i> function transforming the hash value into a
sequence of hash values, and a <i>range-hashing</i> function
transforming the sequence of hash values into a sequence of
positions.</p>
<h3><a name="range_hashing_fns" id=
"range_hashing_fns">Range-Hashing Functions</a></h3>
<p>Some common choices for range-hashing functions are the
division, multiplication, and middle-square methods [<a href=
"references.html#knuth98sorting">knuth98sorting</a>], defined
as</p>
<p><i><a name="division_method" id="division_method">g(r, m) =
r mod m</a></i> (2) ,</p>
<p><i>g(r, m) = ⌈ u/v ( a r mod v ) ⌉</i> ,</p>
<p>and</p>
<p><i>g(r, m) = ⌈ u/v ( r<sup>2</sup> mod v ) ⌉</i>
,</p>
<p>respectively, for some positive integrals <i>u</i> and
<i>v</i> (typically powers of 2), and some <i>a</i>. Each of
these range-hashing functions works best for some different
setting.</p>
<p>The division method <a href="#division_method">(2)</a> is a
very common choice. However, even this single method can be
implemented in two very different ways. It is possible to
implement <a href="#division_method">(2)</a> using the low
level <i>%</i> (modulo) operation (for any <i>m</i>), or the
low level <i>&</i> (bit-mask) operation (for the case where
<i>m</i> is a power of 2), <i>i.e.</i>,</p>
<p><i><a name="division_method_prime_mod" id=
"division_method_prime_mod">g(r, m) = r % m</a></i> (3) ,</p>
<p>and</p>
<p><i><a name="division_method_bit_mask" id=
"division_method_bit_mask">g(r, m) = r & m - 1, (m =
2<sup>k</sup>)</a></i> for some <i>k)</i> (4),</p>
<p>respectively.</p>
<p>The <i>%</i> (modulo) implementation <a href=
"#division_method_prime_mod">(3)</a> has the advantage that for
<i>m</i> a prime far from a power of 2, <i>g(r, m)</i> is
affected by all the bits of <i>r</i> (minimizing the chance of
collision). It has the disadvantage of using the costly modulo
operation. This method is hard-wired into SGI's implementation
[<a href="references.html#sgi_stl">sgi_stl</a>].</p>
<p>The <i>&</i> (bit-mask) implementation <a href=
"#division_method_bit_mask">(4)</a> has the advantage of
relying on the fast bit-wise and operation. It has the
disadvantage that for <i>g(r, m)</i> is affected only by the
low order bits of <i>r</i>. This method is hard-wired into
Dinkumware's implementation [<a href=
"references.html#dinkumware_stl">dinkumware_stl</a>].</p>
<h3><a name="hash_policies_ranged_hash_policies" id=
"hash_policies_ranged_hash_policies">Ranged-Hash
Functions</a></h3>
<p>In cases it is beneficial to allow the
client to directly specify a ranged-hash hash function. It is
true, that the writer of the ranged-hash function cannot rely
on the values of <i>m</i> having specific numerical properties
suitable for hashing (in the sense used in [<a href=
"references.html#knuth98sorting">knuth98sorting</a>]), since
the values of <i>m</i> are determined by a resize policy with
possibly orthogonal considerations.</p>
<p>There are two cases where a ranged-hash function can be
superior. The firs is when using perfect hashing [<a href=
"references.html#knuth98sorting">knuth98sorting</a>]; the
second is when the values of <i>m</i> can be used to estimate
the "general" number of distinct values required. This is
described in the following.</p>
<p>Let</p>
<p class="c2">s = [ s<sub>0</sub>,..., s<sub>t - 1</sub>]</p>
<p>be a string of <i>t</i> characters, each of which is from
domain <i>S</i>. Consider the following ranged-hash
function:</p>
<p><a name="total_string_dna_hash" id=
"total_string_dna_hash"><i>f<sub>1</sub>(s, m) = ∑ <sub>i =
0</sub><sup>t - 1</sup> s<sub>i</sub> a<sup>i</sup></i> mod
<i>m</i></a> (5) ,</p>
<p>where <i>a</i> is some non-negative integral value. This is
the standard string-hashing function used in SGI's
implementation (with <i>a = 5</i>) [<a href=
"references.html#sgi_stl">sgi_stl</a>]. Its advantage is that
it takes into account all of the characters of the string.</p>
<p>Now assume that <i>s</i> is the string representation of a
of a long DNA sequence (and so <i>S = {'A', 'C', 'G',
'T'}</i>). In this case, scanning the entire string might be
prohibitively expensive. A possible alternative might be to use
only the first <i>k</i> characters of the string, where</p>
<p>|S|<sup>k</sup> ≥ m ,</p>
<p><i>i.e.</i>, using the hash function</p>
<p><a name="only_k_string_dna_hash" id=
"only_k_string_dna_hash"><i>f<sub>2</sub>(s, m) = ∑ <sub>i
= 0</sub><sup>k - 1</sup> s<sub>i</sub> a<sup>i</sup></i> mod
<i>m</i></a> , (6)</p>
<p>requiring scanning over only</p>
<p><i>k =</i> log<i><sub>4</sub>( m )</i></p>
<p>characters.</p>
<p>Other more elaborate hash-functions might scan <i>k</i>
characters starting at a random position (determined at each
resize), or scanning <i>k</i> random positions (determined at
each resize), <i>i.e.</i>, using</p>
<p><i>f<sub>3</sub>(s, m) = ∑ <sub>i =
r</sub>0</i><sup>r<sub>0</sub> + k - 1</sup> s<sub>i</sub>
a<sup>i</sup> mod <i>m</i> ,</p>
<p>or</p>
<p><i>f<sub>4</sub>(s, m) = ∑ <sub>i = 0</sub><sup>k -
1</sup> s<sub>r</sub>i</i> a<sup>r<sub>i</sub></sup> mod
<i>m</i> ,</p>
<p>respectively, for <i>r<sub>0</sub>,..., r<sub>k-1</sub></i>
each in the (inclusive) range <i>[0,...,t-1]</i>.</p>
<p>It should be noted that the above functions cannot be
decomposed as <a href=
"#ranged_hash_composed_of_hash_and_range_hashing">(1)</a> .</p>
<h3><a name="pb_ds_imp" id="pb_ds_imp">Implementation</a></h3>
<p>This sub-subsection describes the implementation of the
above in <tt>pb_ds</tt>. It first explains range-hashing
functions in collision-chaining tables, then ranged-hash
functions in collision-chaining tables, then probing-based
tables, and, finally, lists the relevant classes in
<tt>pb_ds</tt>.</p>
<h4>Range-Hashing and Ranged-Hashes in Collision-Chaining
Tables</h4>
<p><a href=
"cc_hash_table.html"><tt>cc_hash_table</tt></a> is
parametrized by <tt>Hash_Fn</tt> and <tt>Comb_Hash_Fn</tt>, a
hash functor and a combining hash functor, respectively.</p>
<p>In general, <tt>Comb_Hash_Fn</tt> is considered a
range-hashing functor. <a href=
"cc_hash_table.html"><tt>cc_hash_table</tt></a>
synthesizes a ranged-hash function from <tt>Hash_Fn</tt> and
<tt>Comb_Hash_Fn</tt> (see <a href=
"#ranged_hash_composed_of_hash_and_range_hashing">(1)</a>
above). Figure <a href="#hash_range_hashing_seq_diagram">Insert
hash sequence diagram</a> shows an <tt>insert</tt> sequence
diagram for this case. The user inserts an element (point A),
the container transforms the key into a non-negative integral
using the hash functor (points B and C), and transforms the
result into a position using the combining functor (points D
and E).</p>
<h6 class="c1"><a name="hash_range_hashing_seq_diagram" id=
"hash_range_hashing_seq_diagram"><img src=
"hash_range_hashing_seq_diagram.png" alt="no image" /></a></h6>
<h6 class="c1">Insert hash sequence diagram.</h6>
<p>If <a href=
"cc_hash_table.html"><tt>cc_hash_table</tt></a>'s
hash-functor, <tt>Hash_Fn</tt> is instantiated by <a href=
"null_hash_fn.html"><tt>null_hash_fn</tt></a> (see <a href=
"concepts.html#concepts_null_policies">Interface::Concepts::Null
Policy Classes</a>), then <tt>Comb_Hash_Fn</tt> is taken to be
a ranged-hash function. Figure <a href=
"#hash_range_hashing_seq_diagram2">Insert hash sequence diagram
with a null hash policy</a> shows an <tt>insert</tt> sequence
diagram. The user inserts an element (point A), the container
transforms the key into a position using the combining functor
(points B and C).</p>
<h6 class="c1"><a name="hash_range_hashing_seq_diagram2" id=
"hash_range_hashing_seq_diagram2"><img src=
"hash_range_hashing_seq_diagram2.png" alt=
"no image" /></a></h6>
<h6 class="c1">Insert hash sequence diagram with a null hash
policy.</h6>
<h4>Probing Tables</h4>
<p><a href=
"gp_hash_table.html"></a><tt>gp_hash_table</tt> is
parametrized by <tt>Hash_Fn</tt>, <tt>Probe_Fn</tt>, and
<tt>Comb_Probe_Fn</tt>. As before, if <tt>Hash_Fn</tt> and
<tt>Probe_Fn</tt> are, respectively, <a href=
"null_hash_fn.html"><tt>null_hash_fn</tt></a> and <a href=
"null_probe_fn.html"><tt>null_probe_fn</tt></a>, then
<tt>Comb_Probe_Fn</tt> is a ranged-probe functor. Otherwise,
<tt>Hash_Fn</tt> is a hash functor, <tt>Probe_Fn</tt> is a
functor for offsets from a hash value, and
<tt>Comb_Probe_Fn</tt> transforms a probe sequence into a
sequence of positions within the table.</p>
<h4>Pre-Defined Policies</h4>
<p><tt>pb_ds</tt> contains some pre-defined classes
implementing range-hashing and probing functions:</p>
<ol>
<li><a href=
"direct_mask_range_hashing.html"><tt>direct_mask_range_hashing</tt></a>
and <a href=
"direct_mod_range_hashing.html"><tt>direct_mod_range_hashing</tt></a>
are range-hashing functions based on a bit-mask and a modulo
operation, respectively.</li>
<li><a href=
"linear_probe_fn.html"><tt>linear_probe_fn</tt></a>, and
<a href=
"quadratic_probe_fn.html"><tt>quadratic_probe_fn</tt></a> are
a linear probe and a quadratic probe function,
respectively.</li>
</ol>Figure <a href="#hash_policy_cd">Hash policy class
diagram</a> shows a class diagram.
<h6 class="c1"><a name="hash_policy_cd" id=
"hash_policy_cd"><img src="hash_policy_cd.png" alt=
"no image" /></a></h6>
<h6 class="c1">Hash policy class diagram.</h6>
<h2><a name="resize_policies" id="resize_policies">Resize
Policies</a></h2>
<h3><a name="general" id="general">General Terms</a></h3>
<p>Hash-tables, as opposed to trees, do not naturally grow or
shrink. It is necessary to specify policies to determine how
and when a hash table should change its size. Usually, resize
policies can be decomposed into orthogonal policies:</p>
<ol>
<li>A <i>size policy</i> indicating <i>how</i> a hash table
should grow (<i>e.g.,</i> it should multiply by powers of
2).</li>
<li>A <i>trigger policy</i> indicating <i>when</i> a hash
table should grow (<i>e.g.,</i> a load factor is
exceeded).</li>
</ol>
<h3><a name="size_policies" id="size_policies">Size
Policies</a></h3>
<p>Size policies determine how a hash table changes size. These
policies are simple, and there are relatively few sensible
options. An exponential-size policy (with the initial size and
growth factors both powers of 2) works well with a mask-based
range-hashing function (see <a href=
"#hash_policies">Range-Hashing Policies</a>), and is the
hard-wired policy used by Dinkumware [<a href=
"references.html#dinkumware_stl">dinkumware_stl</a>]. A
prime-list based policy works well with a modulo-prime range
hashing function (see <a href="#hash_policies">Range-Hashing
Policies</a>), and is the hard-wired policy used by SGI's
implementation [<a href=
"references.html#sgi_stl">sgi_stl</a>].</p>
<h3><a name="trigger_policies" id="trigger_policies">Trigger
Policies</a></h3>
<p>Trigger policies determine when a hash table changes size.
Following is a description of two policies: <i>load-check</i>
policies, and collision-check policies.</p>
<p>Load-check policies are straightforward. The user specifies
two factors, <i>α<sub>min</sub></i> and
<i>α<sub>max</sub></i>, and the hash table maintains the
invariant that</p>
<p><i><a name="load_factor_min_max" id=
"load_factor_min_max">α<sub>min</sub> ≤ (number of
stored elements) / (hash-table size) ≤
α<sub>max</sub></a></i> (1) .</p>
<p>Collision-check policies work in the opposite direction of
load-check policies. They focus on keeping the number of
collisions moderate and hoping that the size of the table will
not grow very large, instead of keeping a moderate load-factor
and hoping that the number of collisions will be small. A
maximal collision-check policy resizes when the longest
probe-sequence grows too large.</p>
<p>Consider Figure <a href="#balls_and_bins">Balls and
bins</a>. Let the size of the hash table be denoted by
<i>m</i>, the length of a probe sequence be denoted by
<i>k</i>, and some load factor be denoted by α. We would
like to calculate the minimal length of <i>k</i>, such that if
there were <i>α m</i> elements in the hash table, a probe
sequence of length <i>k</i> would be found with probability at
most <i>1/m</i>.</p>
<h6 class="c1"><a name="balls_and_bins" id=
"balls_and_bins"><img src="balls_and_bins.png" alt=
"no image" /></a></h6>
<h6 class="c1">Balls and bins.</h6>
<p>Denote the probability that a probe sequence of length
<i>k</i> appears in bin <i>i</i> by <i>p<sub>i</sub></i>, the
length of the probe sequence of bin <i>i</i> by
<i>l<sub>i</sub></i>, and assume uniform distribution. Then</p>
<p><a name="prob_of_p1" id=
"prob_of_p1"><i>p<sub>1</sub></i></a> = (3)</p>
<p class="c2"><b>P</b>(l<sub>1</sub> ≥ k) =</p>
<p><i><b>P</b>(l<sub>1</sub> ≥ α ( 1 + k / α - 1
) ≤</i> (a)</p>
<p><i>e ^ ( - ( α ( k / α - 1 )<sup>2</sup> ) /2
)</i> ,</p>
<p>where (a) follows from the Chernoff bound [<a href=
"references.html#motwani95random">motwani95random</a>]. To
calculate the probability that <i>some</i> bin contains a probe
sequence greater than <i>k</i>, we note that the
<i>l<sub>i</sub></i> are negatively-dependent [<a href=
"references.html#dubhashi98neg">dubhashi98neg</a>]. Let
<i><b>I</b>(.)</i> denote the indicator function. Then</p>
<p><a name="at_least_k_i_n_some_bin" id=
"at_least_k_i_n_some_bin"><i><b>P</b>( exists<sub>i</sub>
l<sub>i</sub> ≥ k ) =</i> (3)</a></p>
<p class="c2"><b>P</b> ( ∑ <sub>i = 1</sub><sup>m</sup>
<b>I</b>(l<sub>i</sub> ≥ k) ≥ 1 ) =</p>
<p><i><b>P</b> ( ∑ <sub>i = 1</sub><sup>m</sup> <b>I</b> (
l<sub>i</sub> ≥ k ) ≥ m p<sub>1</sub> ( 1 + 1 / (m
p<sub>1</sub>) - 1 ) ) ≤</i> (a)</p>
<p class="c2">e ^ ( ( - m p<sub>1</sub> ( 1 / (m p<sub>1</sub>)
- 1 ) <sup>2</sup> ) / 2 ) ,</p>
<p>where (a) follows from the fact that the Chernoff bound can
be applied to negatively-dependent variables [<a href=
"references.html#dubhashi98neg">dubhashi98neg</a>]. Inserting
<a href="#prob_of_p1">(2)</a> into <a href=
"#at_least_k_i_n_some_bin">(3)</a>, and equating with
<i>1/m</i>, we obtain</p>
<p><i>k ~ √ ( 2 α</i> ln <i>2 m</i> ln<i>(m) )
)</i> .</p>
<h3><a name="imp_pb_ds" id="imp_pb_ds">Implementation</a></h3>
<p>This sub-subsection describes the implementation of the
above in <tt>pb_ds</tt>. It first describes resize policies and
their decomposition into trigger and size policies, then
describes pre-defined classes, and finally discusses controlled
access the policies' internals.</p>
<h4>Resize Policies and Their Decomposition</h4>
<p>Each hash-based container is parametrized by a
<tt>Resize_Policy</tt> parameter; the container derives
<tt><b>public</b></tt>ly from <tt>Resize_Policy</tt>. For
example:</p>
<pre>
<a href="cc_hash_table.html">cc_hash_table</a><
<b>typename</b> Key,
<b>typename</b> Mapped,
...
<b>typename</b> Resize_Policy
...> :
<b>public</b> Resize_Policy
</pre>
<p>As a container object is modified, it continuously notifies
its <tt>Resize_Policy</tt> base of internal changes
(<i>e.g.</i>, collisions encountered and elements being
inserted). It queries its <tt>Resize_Policy</tt> base whether
it needs to be resized, and if so, to what size.</p>
<p>Figure <a href="#insert_resize_sequence_diagram1">Insert
resize sequence diagram</a> shows a (possible) sequence diagram
of an insert operation. The user inserts an element; the hash
table notifies its resize policy that a search has started
(point A); in this case, a single collision is encountered -
the table notifies its resize policy of this (point B); the
container finally notifies its resize policy that the search
has ended (point C); it then queries its resize policy whether
a resize is needed, and if so, what is the new size (points D
to G); following the resize, it notifies the policy that a
resize has completed (point H); finally, the element is
inserted, and the policy notified (point I).</p>
<h6 class="c1"><a name="insert_resize_sequence_diagram1" id=
"insert_resize_sequence_diagram1"><img src=
"insert_resize_sequence_diagram1.png" alt=
"no image" /></a></h6>
<h6 class="c1">Insert resize sequence diagram.</h6>
<p>In practice, a resize policy can be usually orthogonally
decomposed to a size policy and a trigger policy. Consequently,
the library contains a single class for instantiating a resize
policy: <a href=
"hash_standard_resize_policy.html"><tt>hash_standard_resize_policy</tt></a>
is parametrized by <tt>Size_Policy</tt> and
<tt>Trigger_Policy</tt>, derives <tt><b>public</b></tt>ly from
both, and acts as a standard delegate [<a href=
"references.html#gamma95designpatterns">gamma95designpatterns</a>]
to these policies.</p>
<p>Figures <a href="#insert_resize_sequence_diagram2">Standard
resize policy trigger sequence diagram</a> and <a href=
"#insert_resize_sequence_diagram3">Standard resize policy size
sequence diagram</a> show sequence diagrams illustrating the
interaction between the standard resize policy and its trigger
and size policies, respectively.</p>
<h6 class="c1"><a name="insert_resize_sequence_diagram2" id=
"insert_resize_sequence_diagram2"><img src=
"insert_resize_sequence_diagram2.png" alt=
"no image" /></a></h6>
<h6 class="c1">Standard resize policy trigger sequence
diagram.</h6>
<h6 class="c1"><a name="insert_resize_sequence_diagram3" id=
"insert_resize_sequence_diagram3"><img src=
"insert_resize_sequence_diagram3.png" alt=
"no image" /></a></h6>
<h6 class="c1">Standard resize policy size sequence
diagram.</h6>
<h4>Pre-Defined Policies</h4>
<p>The library includes the following
instantiations of size and trigger policies:</p>
<ol>
<li><a href=
"hash_load_check_resize_trigger.html"><tt>hash_load_check_resize_trigger</tt></a>
implements a load check trigger policy.</li>
<li><a href=
"cc_hash_max_collision_check_resize_trigger.html"><tt>cc_hash_max_collision_check_resize_trigger</tt></a>
implements a collision check trigger policy.</li>
<li><a href=
"hash_exponential_size_policy.html"><tt>hash_exponential_size_policy</tt></a>
implements an exponential-size policy (which should be used
with mask range hashing).</li>
<li><a href=
"hash_prime_size_policy.html"><tt>hash_prime_size_policy</tt></a>
implementing a size policy based on a sequence of primes
[<a href="references.html#sgi_stl">sgi_stl</a>] (which should
be used with mod range hashing</li>
</ol>
<p>Figure <a href="#resize_policy_cd">Resize policy class
diagram</a> gives an overall picture of the resize-related
classes. <a href=
"basic_hash_table.html"><tt>basic_hash_table</tt></a>
is parametrized by <tt>Resize_Policy</tt>, which it subclasses
publicly. This class is currently instantiated only by <a href=
"hash_standard_resize_policy.html"><tt>hash_standard_resize_policy</tt></a>.
<a href=
"hash_standard_resize_policy.html"><tt>hash_standard_resize_policy</tt></a>
itself is parametrized by <tt>Trigger_Policy</tt> and
<tt>Size_Policy</tt>. Currently, <tt>Trigger_Policy</tt> is
instantiated by <a href=
"hash_load_check_resize_trigger.html"><tt>hash_load_check_resize_trigger</tt></a>,
or <a href=
"cc_hash_max_collision_check_resize_trigger.html"><tt>cc_hash_max_collision_check_resize_trigger</tt></a>;
<tt>Size_Policy</tt> is instantiated by <a href=
"hash_exponential_size_policy.html"><tt>hash_exponential_size_policy</tt></a>,
or <a href=
"hash_prime_size_policy.html"><tt>hash_prime_size_policy</tt></a>.</p>
<h6 class="c1"><a name="resize_policy_cd" id=
"resize_policy_cd"><img src="resize_policy_cd.png" alt=
"no image" /></a></h6>
<h6 class="c1">Resize policy class diagram.</h6>
<h4>Controlled Access to Policies' Internals</h4>
<p>There are cases where (controlled) access to resize
policies' internals is beneficial. <i>E.g.</i>, it is sometimes
useful to query a hash-table for the table's actual size (as
opposed to its <tt>size()</tt> - the number of values it
currently holds); it is sometimes useful to set a table's
initial size, externally resize it, or change load factors.</p>
<p>Clearly, supporting such methods both decreases the
encapsulation of hash-based containers, and increases the
diversity between different associative-containers' interfaces.
Conversely, omitting such methods can decrease containers'
flexibility.</p>
<p>In order to avoid, to the extent possible, the above
conflict, the hash-based containers themselves do not address
any of these questions; this is deferred to the resize policies,
which are easier to change or replace. Thus, for example,
neither <a href=
"cc_hash_table.html"><tt>cc_hash_table</tt></a> nor
<a href=
"gp_hash_table.html"><tt>gp_hash_table</tt></a>
contain methods for querying the actual size of the table; this
is deferred to <a href=
"hash_standard_resize_policy.html"><tt>hash_standard_resize_policy</tt></a>.</p>
<p>Furthermore, the policies themselves are parametrized by
template arguments that determine the methods they support
([<a href=
"references.html#alexandrescu01modern">alexandrescu01modern</a>]
shows techniques for doing so). <a href=
"hash_standard_resize_policy.html"><tt>hash_standard_resize_policy</tt></a>
is parametrized by <tt>External_Size_Access</tt> that
determines whether it supports methods for querying the actual
size of the table or resizing it. <a href=
"hash_load_check_resize_trigger.html"><tt>hash_load_check_resize_trigger</tt></a>
is parametrized by <tt>External_Load_Access</tt> that
determines whether it supports methods for querying or
modifying the loads. <a href=
"cc_hash_max_collision_check_resize_trigger.html"><tt>cc_hash_max_collision_check_resize_trigger</tt></a>
is parametrized by <tt>External_Load_Access</tt> that
determines whether it supports methods for querying the
load.</p>
<p>Some operations, for example, resizing a container at
run time, or changing the load factors of a load-check trigger
policy, require the container itself to resize. As mentioned
above, the hash-based containers themselves do not contain
these types of methods, only their resize policies.
Consequently, there must be some mechanism for a resize policy
to manipulate the hash-based container. As the hash-based
container is a subclass of the resize policy, this is done
through virtual methods. Each hash-based container has a
<tt><b>private</b></tt> <tt><b>virtual</b></tt> method:</p>
<pre>
<b>virtual void</b>
do_resize
(size_type new_size);
</pre>
<p>which resizes the container. Implementations of
<tt>Resize_Policy</tt> can export public methods for resizing
the container externally; these methods internally call
<tt>do_resize</tt> to resize the table.</p>
<h2><a name="policy_interaction" id="policy_interaction">Policy
Interaction</a></h2>
<p>Hash-tables are unfortunately especially susceptible to
choice of policies. One of the more complicated aspects of this
is that poor combinations of good policies can form a poor
container. Following are some considerations.</p>
<h3><a name="policy_interaction_probe_size_trigger" id=
"policy_interaction_probe_size_trigger">Probe Policies, Size
Policies, and Trigger Policies</a></h3>
<p>Some combinations do not work well for probing containers.
For example, combining a quadratic probe policy with an
exponential size policy can yield a poor container: when an
element is inserted, a trigger policy might decide that there
is no need to resize, as the table still contains unused
entries; the probe sequence, however, might never reach any of
the unused entries.</p>
<p>Unfortunately, <tt>pb_ds</tt> cannot detect such problems at
compilation (they are halting reducible). It therefore defines
an exception class <a href=
"insert_error.html"><tt>insert_error</tt></a> to throw an
exception in this case.</p>
<h3><a name="policy_interaction_hash_trigger" id=
"policy_interaction_hash_trigger">Hash Policies and Trigger
Policies</a></h3>
<p>Some trigger policies are especially susceptible to poor
hash functions. Suppose, as an extreme case, that the hash
function transforms each key to the same hash value. After some
inserts, a collision detecting policy will always indicate that
the container needs to grow.</p>
<p>The library, therefore, by design, limits each operation to
one resize. For each <tt>insert</tt>, for example, it queries
only once whether a resize is needed.</p>
<h3><a name="policy_interaction_eq_sth_hash" id=
"policy_interaction_eq_sth_hash">Equivalence Functors, Storing
Hash Values, and Hash Functions</a></h3>
<p><a href=
"cc_hash_table.html"><tt>cc_hash_table</tt></a> and
<a href=
"gp_hash_table.html"><tt>gp_hash_table</tt></a> are
parametrized by an equivalence functor and by a
<tt>Store_Hash</tt> parameter. If the latter parameter is
<tt><b>true</b></tt>, then the container stores with each entry
a hash value, and uses this value in case of collisions to
determine whether to apply a hash value. This can lower the
cost of collision for some types, but increase the cost of
collisions for other types.</p>
<p>If a ranged-hash function or ranged probe function is
directly supplied, however, then it makes no sense to store the
hash value with each entry. <tt>pb_ds</tt>'s container will
fail at compilation, by design, if this is attempted.</p>
<h3><a name="policy_interaction_size_load_check" id=
"policy_interaction_size_load_check">Size Policies and
Load-Check Trigger Policies</a></h3>
<p>Assume a size policy issues an increasing sequence of sizes
<i>a, a q, a q<sup>1</sup>, a q<sup>2</sup>, ...</i> For
example, an exponential size policy might issue the sequence of
sizes <i>8, 16, 32, 64, ...</i></p>
<p>If a load-check trigger policy is used, with loads
<i>α<sub>min</sub></i> and <i>α<sub>max</sub></i>,
respectively, then it is a good idea to have:</p>
<ol>
<li><i>α<sub>max</sub> ~ 1 / q</i></li>
<li><i>α<sub>min</sub> < 1 / (2 q)</i></li>
</ol>
<p>This will ensure that the amortized hash cost of each
modifying operation is at most approximately 3.</p>
<p><i>α<sub>min</sub> ~ α<sub>max</sub></i> is, in
any case, a bad choice, and <i>α<sub>min</sub> >
α<sub>max</sub></i> is horrendous.</p>
</div>
</body>
</html>
