1 // Copyright (c) 2005, Google Inc.
2 // All rights reserved.
4 // Redistribution and use in source and binary forms, with or without
5 // modification, are permitted provided that the following conditions are
8 // * Redistributions of source code must retain the above copyright
9 // notice, this list of conditions and the following disclaimer.
10 // * Redistributions in binary form must reproduce the above
11 // copyright notice, this list of conditions and the following disclaimer
12 // in the documentation and/or other materials provided with the
14 // * Neither the name of Google Inc. nor the names of its
15 // contributors may be used to endorse or promote products derived from
16 // this software without specific prior written permission.
18 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
19 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
20 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
21 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
22 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
23 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
24 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
25 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
26 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
27 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
28 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31 // Author: Craig Silverstein
33 // This is just a very thin wrapper over sparsehashtable.h, just
34 // like sgi stl's stl_hash_set is a very thin wrapper over
35 // stl_hashtable. The major thing we define is operator[], because
36 // we have a concept of a data_type which stl_hashtable doesn't
37 // (it only has a key and a value).
39 // This is more different from sparse_hash_map than you might think,
40 // because all iterators for sets are const (you obviously can't
41 // change the key, and for sets there is no value).
43 // We adhere mostly to the STL semantics for hash-set. One important
44 // exception is that insert() invalidates iterators entirely. On the
45 // plus side, though, delete() doesn't invalidate iterators at all, or
46 // even change the ordering of elements.
48 // Here are a few "power user" tips:
50 // 1) set_deleted_key():
51 // Unlike STL's hash_map, if you want to use erase() you
52 // must call set_deleted_key() after construction.
55 // When an item is deleted, its memory isn't freed right
56 // away. This allows you to iterate over a hashtable,
57 // and call erase(), without invalidating the iterator.
58 // To force the memory to be freed, call resize(0).
60 // Guide to what kind of hash_set to use:
61 // (1) dense_hash_set: fastest, uses the most memory
62 // (2) sparse_hash_set: slowest, uses the least memory
63 // (3) hash_set (STL): in the middle
64 // Typically I use sparse_hash_set when I care about space and/or when
65 // I need to save the hashtable on disk. I use hash_set otherwise. I
66 // don't personally use dense_hash_set ever; the only use of
67 // dense_hash_set I know of is to work around malloc() bugs in some
68 // systems (dense_hash_set has a particularly simple allocation scheme).
70 // - dense_hash_set has, typically, a factor of 2 memory overhead (if your
71 // data takes up X bytes, the hash_set uses X more bytes in overhead).
72 // - sparse_hash_set has about 2 bits overhead per entry.
73 // - sparse_hash_map can be 3-7 times slower than the others for lookup and,
74 // especially, inserts. See time_hash_map.cc for details.
76 // See /usr/(local/)?doc/sparsehash-0.1/sparse_hash_set.html
77 // for information about how to use this class.
79 #ifndef _SPARSE_HASH_SET_H_
80 #define _SPARSE_HASH_SET_H_
82 // !!! DR changed some <google/...> to "..." due to include path problems...
83 #include "google_config.h"
84 #include <stdio.h> // for FILE * in read()/write()
85 #include <algorithm> // for the default template args
86 #include <functional> // for equal_to
87 #include <memory> // for alloc<>
89 #include "sparsehashtable.h"
91 _START_GOOGLE_NAMESPACE_
93 using STL_NAMESPACE::pair;
95 template <class Value,
96 class HashFcn = HASH_NAMESPACE::hash<Value>,
97 class EqualKey = STL_NAMESPACE::equal_to<Value>,
98 class Alloc = STL_NAMESPACE::allocator<Value> >
99 class sparse_hash_set {
102 // Apparently identity is not stl-standard, so we define our own
104 Value& operator()(Value& v) const { return v; }
105 const Value& operator()(const Value& v) const { return v; }
109 typedef sparse_hashtable<Value, Value, HashFcn, Identity, EqualKey, Alloc> ht;
113 typedef typename ht::key_type key_type;
114 typedef typename ht::value_type value_type;
115 typedef typename ht::hasher hasher;
116 typedef typename ht::key_equal key_equal;
118 typedef typename ht::size_type size_type;
119 typedef typename ht::difference_type difference_type;
120 typedef typename ht::const_pointer pointer;
121 typedef typename ht::const_pointer const_pointer;
122 typedef typename ht::const_reference reference;
123 typedef typename ht::const_reference const_reference;
125 typedef typename ht::const_iterator iterator;
126 typedef typename ht::const_iterator const_iterator;
129 // Iterator functions -- recall all iterators are const
130 iterator begin() const { return rep.begin(); }
131 iterator end() const { return rep.end(); }
134 // Accessor functions
135 hasher hash_funct() const { return rep.hash_funct(); }
136 key_equal key_eq() const { return rep.key_eq(); }
140 explicit sparse_hash_set(size_type n = 0,
141 const hasher& hf = hasher(),
142 const key_equal& eql = key_equal())
143 : rep(n, hf, eql) { }
145 template <class InputIterator>
146 sparse_hash_set(InputIterator f, InputIterator l,
148 const hasher& hf = hasher(),
149 const key_equal& eql = key_equal()) {
152 // We use the default copy constructor
153 // We use the default operator=()
154 // We use the default destructor
156 void clear() { rep.clear(); }
157 void swap(sparse_hash_set& hs) { rep.swap(hs.rep); }
160 // Functions concerning size
161 size_type size() const { return rep.size(); }
162 size_type max_size() const { return rep.max_size(); }
163 bool empty() const { return rep.empty(); }
164 size_type bucket_count() const { return rep.bucket_count(); }
165 size_type max_bucket_count() const { return rep.max_bucket_count(); }
167 void resize(size_type hint) { rep.resize(hint); }
171 iterator find(const key_type& key) const { return rep.find(key); }
173 size_type count(const key_type& key) const { return rep.count(key); }
175 pair<iterator, iterator> equal_range(const key_type& key) const {
176 return rep.equal_range(key);
179 // Insertion routines
180 pair<iterator, bool> insert(const value_type& obj) {
181 pair<typename ht::iterator, bool> p = rep.insert(obj);
182 return pair<iterator, bool>(p.first, p.second); // const to non-const
184 template <class InputIterator>
185 void insert(InputIterator f, InputIterator l) { rep.insert(f, l); }
186 void insert(const_iterator f, const_iterator l) { rep.insert(f, l); }
187 // required for std::insert_iterator; the passed-in iterator is ignored
188 iterator insert(iterator, const value_type& obj) { return insert(obj).first; }
192 // THESE ARE NON-STANDARD! I make you specify an "impossible" key
193 // value to identify deleted buckets. You can change the key as
194 // time goes on, or get rid of it entirely to be insert-only.
195 void set_deleted_key(const key_type& key) { rep.set_deleted_key(key); }
196 void clear_deleted_key() { rep.clear_deleted_key(); }
198 // These are standard
199 size_type erase(const key_type& key) { return rep.erase(key); }
200 void erase(iterator it) { rep.erase(it); }
201 void erase(iterator f, iterator l) { rep.erase(f, l); }
205 bool operator==(const sparse_hash_set& hs) const { return rep == hs.rep; }
206 bool operator!=(const sparse_hash_set& hs) const { return rep != hs.rep; }
209 // I/O -- this is an add-on for writing metainformation to disk
210 bool write_metadata(FILE *fp) { return rep.write_metadata(fp); }
211 bool read_metadata(FILE *fp) { return rep.read_metadata(fp); }
212 bool write_nopointer_data(FILE *fp) { return rep.write_nopointer_data(fp); }
213 bool read_nopointer_data(FILE *fp) { return rep.read_nopointer_data(fp); }
216 template <class Val, class HashFcn, class EqualKey, class Alloc>
217 inline void swap(sparse_hash_set<Val, HashFcn, EqualKey, Alloc>& hs1,
218 sparse_hash_set<Val, HashFcn, EqualKey, Alloc>& hs2) {
222 _END_GOOGLE_NAMESPACE_
224 #endif /* _SPARSE_HASH_SET_H_ */