1 | /*
|
---|
2 | * Copyright (c) 2018 Jaroslav Jindrak
|
---|
3 | * All rights reserved.
|
---|
4 | *
|
---|
5 | * Redistribution and use in source and binary forms, with or without
|
---|
6 | * modification, are permitted provided that the following conditions
|
---|
7 | * are met:
|
---|
8 | *
|
---|
9 | * - Redistributions of source code must retain the above copyright
|
---|
10 | * notice, this list of conditions and the following disclaimer.
|
---|
11 | * - Redistributions in binary form must reproduce the above copyright
|
---|
12 | * notice, this list of conditions and the following disclaimer in the
|
---|
13 | * documentation and/or other materials provided with the distribution.
|
---|
14 | * - The name of the author may not be used to endorse or promote products
|
---|
15 | * derived from this software without specific prior written permission.
|
---|
16 | *
|
---|
17 | * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
|
---|
18 | * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
|
---|
19 | * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
|
---|
20 | * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
|
---|
21 | * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
|
---|
22 | * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
|
---|
23 | * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
|
---|
24 | * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
|
---|
25 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
|
---|
26 | * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
---|
27 | */
|
---|
28 |
|
---|
29 | #ifndef LIBCPP_BITS_HASH_TABLE
|
---|
30 | #define LIBCPP_BITS_HASH_TABLE
|
---|
31 |
|
---|
32 | #include <cstdlib>
|
---|
33 | #include <__bits/list_node.hpp>
|
---|
34 | #include <__bits/key_extractors.hpp>
|
---|
35 | #include <__bits/hash_table_iterators.hpp>
|
---|
36 | #include <__bits/hash_table_policies.hpp>
|
---|
37 | #include <iterator>
|
---|
38 | #include <limits>
|
---|
39 | #include <memory>
|
---|
40 | #include <tuple>
|
---|
41 | #include <utility>
|
---|
42 |
|
---|
43 | namespace std::aux
|
---|
44 | {
|
---|
45 | /**
|
---|
46 | * To save code, we're going to implement one hash table
|
---|
47 | * for both unordered_map and unordered_set. To do this,
|
---|
48 | * we create one inner hash table that is oblivious to its
|
---|
49 | * held type (and uses a key extractor to get the key from it)
|
---|
50 | * and two proxies that either use plain Key type or a pair
|
---|
51 | * of a key and a value.
|
---|
52 | * Additionally, we will use policies to represent both single
|
---|
53 | * and multi variants of these containers at once.
|
---|
54 | * Note: I am aware the naming is very unimaginative here,
|
---|
55 | * not my strong side :)
|
---|
56 | */
|
---|
57 |
|
---|
58 | template<
|
---|
59 | class Value, class Key, class KeyExtractor,
|
---|
60 | class Hasher, class KeyEq,
|
---|
61 | class Alloc, class Size,
|
---|
62 | class Iterator, class ConstIterator,
|
---|
63 | class LocalIterator, class ConstLocalIterator,
|
---|
64 | class Policy
|
---|
65 | >
|
---|
66 | class hash_table
|
---|
67 | {
|
---|
68 | public:
|
---|
69 | using value_type = Value;
|
---|
70 | using key_type = Key;
|
---|
71 | using size_type = Size;
|
---|
72 | using allocator_type = Alloc;
|
---|
73 | using key_equal = KeyEq;
|
---|
74 | using hasher = Hasher;
|
---|
75 | using key_extract = KeyExtractor;
|
---|
76 |
|
---|
77 | using iterator = Iterator;
|
---|
78 | using const_iterator = ConstIterator;
|
---|
79 | using local_iterator = LocalIterator;
|
---|
80 | using const_local_iterator = ConstLocalIterator;
|
---|
81 |
|
---|
82 | using node_type = list_node<value_type>;
|
---|
83 |
|
---|
84 | using place_type = tuple<
|
---|
85 | hash_table_bucket<value_type, size_type>*,
|
---|
86 | list_node<value_type>*, size_type
|
---|
87 | >;
|
---|
88 |
|
---|
89 | hash_table(size_type buckets, float max_load_factor = 1.f)
|
---|
90 | : table_{new hash_table_bucket<value_type, size_type>[buckets]()},
|
---|
91 | bucket_count_{buckets}, size_{}, hasher_{}, key_eq_{},
|
---|
92 | key_extractor_{}, max_load_factor_{max_load_factor}
|
---|
93 | { /* DUMMY BODY */ }
|
---|
94 |
|
---|
95 | hash_table(size_type buckets, const hasher& hf, const key_equal& eql,
|
---|
96 | float max_load_factor = 1.f)
|
---|
97 | : table_{new hash_table_bucket<value_type, size_type>[buckets]()},
|
---|
98 | bucket_count_{buckets}, size_{}, hasher_{hf}, key_eq_{eql},
|
---|
99 | key_extractor_{}, max_load_factor_{max_load_factor}
|
---|
100 | { /* DUMMY BODY */ }
|
---|
101 |
|
---|
102 | hash_table(const hash_table& other)
|
---|
103 | : hash_table{other.bucket_count_, other.hasher_, other.key_eq_,
|
---|
104 | other.max_load_factor_}
|
---|
105 | {
|
---|
106 | for (const auto& x: other)
|
---|
107 | insert(x);
|
---|
108 | }
|
---|
109 |
|
---|
110 | hash_table(hash_table&& other)
|
---|
111 | : table_{other.table_}, bucket_count_{other.bucket_count_},
|
---|
112 | size_{other.size_}, hasher_{move(other.hasher_)},
|
---|
113 | key_eq_{move(other.key_eq_)}, key_extractor_{move(other.key_extractor_)},
|
---|
114 | max_load_factor_{other.max_load_factor_}
|
---|
115 | {
|
---|
116 | other.table_ = nullptr;
|
---|
117 | other.bucket_count_ = size_type{};
|
---|
118 | other.size_ = size_type{};
|
---|
119 | other.max_load_factor_ = 1.f;
|
---|
120 | }
|
---|
121 |
|
---|
122 | hash_table& operator=(const hash_table& other)
|
---|
123 | {
|
---|
124 | hash_table tmp{other};
|
---|
125 | tmp.swap(*this);
|
---|
126 |
|
---|
127 | return *this;
|
---|
128 | }
|
---|
129 |
|
---|
130 | hash_table& operator=(hash_table&& other)
|
---|
131 | {
|
---|
132 | hash_table tmp{move(other)};
|
---|
133 | tmp.swap(*this);
|
---|
134 |
|
---|
135 | return *this;
|
---|
136 | }
|
---|
137 |
|
---|
138 | bool empty() const noexcept
|
---|
139 | {
|
---|
140 | return size_ == 0;
|
---|
141 | }
|
---|
142 |
|
---|
143 | size_type size() const noexcept
|
---|
144 | {
|
---|
145 | return size_;
|
---|
146 | }
|
---|
147 |
|
---|
148 | size_type max_size(allocator_type& alloc)
|
---|
149 | {
|
---|
150 | return allocator_traits<allocator_type>::max_size(alloc);
|
---|
151 | }
|
---|
152 |
|
---|
153 | iterator begin() noexcept
|
---|
154 | {
|
---|
155 | auto idx = first_filled_bucket_();
|
---|
156 | return iterator{
|
---|
157 | table_, idx, bucket_count_,
|
---|
158 | table_[idx].head
|
---|
159 | };
|
---|
160 | }
|
---|
161 |
|
---|
162 | const_iterator begin() const noexcept
|
---|
163 | {
|
---|
164 | return cbegin();
|
---|
165 | }
|
---|
166 |
|
---|
167 | iterator end() noexcept
|
---|
168 | {
|
---|
169 | return iterator{};
|
---|
170 | }
|
---|
171 |
|
---|
172 | const_iterator end() const noexcept
|
---|
173 | {
|
---|
174 | return cend();
|
---|
175 | }
|
---|
176 |
|
---|
177 | const_iterator cbegin() const noexcept
|
---|
178 | {
|
---|
179 | auto idx = first_filled_bucket_();
|
---|
180 | return const_iterator{
|
---|
181 | table_, idx, bucket_count_,
|
---|
182 | table_[idx].head
|
---|
183 | };
|
---|
184 | }
|
---|
185 |
|
---|
186 | const_iterator cend() const noexcept
|
---|
187 | {
|
---|
188 | return const_iterator{};
|
---|
189 | }
|
---|
190 |
|
---|
191 | template<class... Args>
|
---|
192 | auto emplace(Args&&... args)
|
---|
193 | {
|
---|
194 | return Policy::emplace(*this, forward<Args>(args)...);
|
---|
195 | }
|
---|
196 |
|
---|
197 | auto insert(const value_type& val)
|
---|
198 | {
|
---|
199 | return Policy::insert(*this, val);
|
---|
200 | }
|
---|
201 |
|
---|
202 | auto insert(value_type&& val)
|
---|
203 | {
|
---|
204 | return Policy::insert(*this, forward<value_type>(val));
|
---|
205 | }
|
---|
206 |
|
---|
207 | size_type erase(const key_type& key)
|
---|
208 | {
|
---|
209 | return Policy::erase(*this, key);
|
---|
210 | }
|
---|
211 |
|
---|
212 | iterator erase(const_iterator it)
|
---|
213 | {
|
---|
214 | auto node = it.node();
|
---|
215 | auto idx = it.idx();
|
---|
216 |
|
---|
217 | /**
|
---|
218 | * Note: This way we will continue on the next bucket
|
---|
219 | * if this is the last element in its bucket.
|
---|
220 | */
|
---|
221 | iterator res{table_, idx, size_, node};
|
---|
222 | ++res;
|
---|
223 |
|
---|
224 | if (table_[idx].head == node)
|
---|
225 | {
|
---|
226 | if (node->next != node)
|
---|
227 | table_[idx].head = node->next;
|
---|
228 | else
|
---|
229 | table_[idx].head = nullptr;
|
---|
230 | }
|
---|
231 | --size_;
|
---|
232 |
|
---|
233 | node->unlink();
|
---|
234 | delete node;
|
---|
235 |
|
---|
236 | return res;
|
---|
237 | }
|
---|
238 |
|
---|
239 | void clear() noexcept
|
---|
240 | {
|
---|
241 | for (size_type i = 0; i < bucket_count_; ++i)
|
---|
242 | table_[i].clear();
|
---|
243 | size_ = size_type{};
|
---|
244 | }
|
---|
245 |
|
---|
246 | void swap(hash_table& other)
|
---|
247 | noexcept(allocator_traits<allocator_type>::is_always_equal::value &&
|
---|
248 | noexcept(swap(declval<Hasher&>(), declval<Hasher&>())) &&
|
---|
249 | noexcept(swap(declval<KeyEq&>(), declval<KeyEq&>())))
|
---|
250 | {
|
---|
251 | std::swap(table_, other.table_);
|
---|
252 | std::swap(bucket_count_, other.bucket_count_);
|
---|
253 | std::swap(size_, other.size_);
|
---|
254 | std::swap(hasher_, other.hasher_);
|
---|
255 | std::swap(key_eq_, other.key_eq_);
|
---|
256 | std::swap(max_load_factor_, other.max_load_factor_);
|
---|
257 | }
|
---|
258 |
|
---|
259 | hasher hash_function() const
|
---|
260 | {
|
---|
261 | return hasher_;
|
---|
262 | }
|
---|
263 |
|
---|
264 | key_equal key_eq() const
|
---|
265 | {
|
---|
266 | return key_eq_;
|
---|
267 | }
|
---|
268 |
|
---|
269 | iterator find(const key_type& key)
|
---|
270 | {
|
---|
271 | auto idx = get_bucket_idx_(key);
|
---|
272 | auto head = table_[idx].head;
|
---|
273 |
|
---|
274 | if (!head)
|
---|
275 | return end();
|
---|
276 |
|
---|
277 | auto current = head;
|
---|
278 | do
|
---|
279 | {
|
---|
280 | if (key_eq_(key, key_extractor_(current->value)))
|
---|
281 | return iterator{table_, idx, size_, current};
|
---|
282 | current = current->next;
|
---|
283 | }
|
---|
284 | while (current != head);
|
---|
285 |
|
---|
286 | return end();
|
---|
287 | }
|
---|
288 |
|
---|
289 | const_iterator find(const key_type& key) const
|
---|
290 | {
|
---|
291 | auto idx = get_bucket_idx_(key);
|
---|
292 | auto head = table_[idx].head;
|
---|
293 |
|
---|
294 | if (!head)
|
---|
295 | return end();
|
---|
296 |
|
---|
297 | auto current = head;
|
---|
298 | do
|
---|
299 | {
|
---|
300 | if (key_eq_(key, key_extractor_(current->value)))
|
---|
301 | return iterator{table_, idx, size_, current};
|
---|
302 | current = current->next;
|
---|
303 | }
|
---|
304 | while (current != head);
|
---|
305 |
|
---|
306 | return end();
|
---|
307 | }
|
---|
308 |
|
---|
309 | size_type count(const key_type& key) const
|
---|
310 | {
|
---|
311 | return Policy::count(*this, key);
|
---|
312 | }
|
---|
313 |
|
---|
314 | pair<iterator, iterator> equal_range(const key_type& key)
|
---|
315 | {
|
---|
316 | return Policy::equal_range(*this, key);
|
---|
317 | }
|
---|
318 |
|
---|
319 | pair<const_iterator, const_iterator> equal_range(const key_type& key) const
|
---|
320 | {
|
---|
321 | return Policy::equal_range_const(*this, key);
|
---|
322 | }
|
---|
323 |
|
---|
324 | size_type bucket_count() const noexcept
|
---|
325 | {
|
---|
326 | return bucket_count_;
|
---|
327 | }
|
---|
328 |
|
---|
329 | size_type max_bucket_count() const noexcept
|
---|
330 | {
|
---|
331 | return numeric_limits<size_type>::max() /
|
---|
332 | sizeof(hash_table_bucket<value_type, size_type>);
|
---|
333 | }
|
---|
334 |
|
---|
335 | size_type bucket_size(size_type n) const
|
---|
336 | {
|
---|
337 | return table_[n].size();
|
---|
338 | }
|
---|
339 |
|
---|
340 | size_type bucket(const key_type& key) const
|
---|
341 | {
|
---|
342 | return get_bucket_idx_(key);
|
---|
343 | }
|
---|
344 |
|
---|
345 | local_iterator begin(size_type n)
|
---|
346 | {
|
---|
347 | return local_iterator{table_[n].head, table_[n].head};
|
---|
348 | }
|
---|
349 |
|
---|
350 | const_local_iterator begin(size_type n) const
|
---|
351 | {
|
---|
352 | return cbegin(n);
|
---|
353 | }
|
---|
354 |
|
---|
355 | local_iterator end(size_type n)
|
---|
356 | {
|
---|
357 | return local_iterator{};
|
---|
358 | }
|
---|
359 |
|
---|
360 | const_local_iterator end(size_type n) const
|
---|
361 | {
|
---|
362 | return cend(n);
|
---|
363 | }
|
---|
364 |
|
---|
365 | const_local_iterator cbegin(size_type n) const
|
---|
366 | {
|
---|
367 | return const_local_iterator{table_[n].head, table_[n].head};
|
---|
368 | }
|
---|
369 |
|
---|
370 | const_local_iterator cend(size_type n) const
|
---|
371 | {
|
---|
372 | return const_local_iterator{};
|
---|
373 | }
|
---|
374 |
|
---|
375 | float load_factor() const noexcept
|
---|
376 | {
|
---|
377 | return size_ / static_cast<float>(bucket_count_);
|
---|
378 | }
|
---|
379 |
|
---|
380 | float max_load_factor() const noexcept
|
---|
381 | {
|
---|
382 | return max_load_factor_;
|
---|
383 | }
|
---|
384 |
|
---|
385 | void max_load_factor(float factor)
|
---|
386 | {
|
---|
387 | if (factor > 0.f)
|
---|
388 | max_load_factor_ = factor;
|
---|
389 |
|
---|
390 | rehash_if_needed();
|
---|
391 | }
|
---|
392 |
|
---|
393 | void rehash(size_type count)
|
---|
394 | {
|
---|
395 | if (count < size_ / max_load_factor_)
|
---|
396 | count = size_ / max_load_factor_;
|
---|
397 |
|
---|
398 | /**
|
---|
399 | * Note: If an exception is thrown, there
|
---|
400 | * is no effect. Since this is the only
|
---|
401 | * place where an exception (no mem) can
|
---|
402 | * be thrown and no changes to this have been
|
---|
403 | * made, we're ok.
|
---|
404 | */
|
---|
405 | hash_table new_table{count, max_load_factor_};
|
---|
406 |
|
---|
407 | for (std::size_t i = 0; i < bucket_count_; ++i)
|
---|
408 | {
|
---|
409 | auto head = table_[i].head;
|
---|
410 | if (!head)
|
---|
411 | continue;
|
---|
412 |
|
---|
413 | auto current = head;
|
---|
414 |
|
---|
415 | do
|
---|
416 | {
|
---|
417 | auto next = current->next;
|
---|
418 |
|
---|
419 | current->next = current;
|
---|
420 | current->prev = current;
|
---|
421 |
|
---|
422 | auto where = Policy::find_insertion_spot(
|
---|
423 | new_table, key_extractor_(current->value)
|
---|
424 | );
|
---|
425 |
|
---|
426 | /**
|
---|
427 | * Note: We're rehashing, so we know each
|
---|
428 | * key can be inserted.
|
---|
429 | */
|
---|
430 | auto new_bucket = get<0>(where);
|
---|
431 | auto new_successor = get<1>(where);
|
---|
432 |
|
---|
433 | if (new_successor)
|
---|
434 | new_successor->append(current);
|
---|
435 | else
|
---|
436 | new_bucket->append(current);
|
---|
437 |
|
---|
438 | current = next;
|
---|
439 | } while (current != head);
|
---|
440 |
|
---|
441 | table_[i].head = nullptr;
|
---|
442 | }
|
---|
443 |
|
---|
444 | new_table.size_ = size_;
|
---|
445 | swap(new_table);
|
---|
446 |
|
---|
447 | delete[] new_table.table_;
|
---|
448 | new_table.table_ = nullptr;
|
---|
449 | }
|
---|
450 |
|
---|
451 | void reserve(size_type count)
|
---|
452 | {
|
---|
453 | rehash(count / max_load_factor_ + 1);
|
---|
454 | }
|
---|
455 |
|
---|
456 | bool is_eq_to(const hash_table& other) const
|
---|
457 | {
|
---|
458 | if (size() != other.size())
|
---|
459 | return false;
|
---|
460 |
|
---|
461 | auto it = begin();
|
---|
462 | while (it != end())
|
---|
463 | {
|
---|
464 | /**
|
---|
465 | * For each key K we will check how many
|
---|
466 | * instances of K are there in the table.
|
---|
467 | * Then we will check if the count for K
|
---|
468 | * is equal to that amount.
|
---|
469 | */
|
---|
470 |
|
---|
471 | size_type cnt{};
|
---|
472 | auto tmp = it;
|
---|
473 |
|
---|
474 | while (key_eq_(key_extractor_(*it), key_extractor_(*tmp)))
|
---|
475 | {
|
---|
476 | ++cnt;
|
---|
477 | if (++tmp == end())
|
---|
478 | break;
|
---|
479 | }
|
---|
480 |
|
---|
481 | auto other_cnt = other.count(key_extractor_(*it));
|
---|
482 | if (cnt != other_cnt)
|
---|
483 | return false;
|
---|
484 |
|
---|
485 | it = tmp; // tmp is one past *it's key.
|
---|
486 | }
|
---|
487 |
|
---|
488 | return true;
|
---|
489 | }
|
---|
490 |
|
---|
491 | ~hash_table()
|
---|
492 | {
|
---|
493 | // Lists are deleted in ~hash_table_bucket.
|
---|
494 | if (table_)
|
---|
495 | delete[] table_;
|
---|
496 | }
|
---|
497 |
|
---|
498 | place_type find_insertion_spot(const key_type& key) const
|
---|
499 | {
|
---|
500 | return Policy::find_insertion_spot(*this, key);
|
---|
501 | }
|
---|
502 |
|
---|
503 | place_type find_insertion_spot(key_type&& key) const
|
---|
504 | {
|
---|
505 | return Policy::find_insertion_spot(*this, key);
|
---|
506 | }
|
---|
507 |
|
---|
508 | const key_type& get_key(const value_type& val) const
|
---|
509 | {
|
---|
510 | return key_extractor_(val);
|
---|
511 | }
|
---|
512 |
|
---|
513 | bool keys_equal(const key_type& key, const value_type& val)
|
---|
514 | {
|
---|
515 | return key_eq_(key, key_extractor_(val));
|
---|
516 | }
|
---|
517 |
|
---|
518 | bool keys_equal(const key_type& key, const value_type& val) const
|
---|
519 | {
|
---|
520 | return key_eq_(key, key_extractor_(val));
|
---|
521 | }
|
---|
522 |
|
---|
523 | hash_table_bucket<value_type, size_type>* table()
|
---|
524 | {
|
---|
525 | return table_;
|
---|
526 | }
|
---|
527 |
|
---|
528 | hash_table_bucket<value_type, size_type>* head(size_type idx)
|
---|
529 | {
|
---|
530 | if (idx < bucket_count_)
|
---|
531 | return table_[idx]->head;
|
---|
532 | else
|
---|
533 | return nullptr;
|
---|
534 | }
|
---|
535 |
|
---|
536 | void rehash_if_needed()
|
---|
537 | {
|
---|
538 | if (size_ > max_load_factor_ * bucket_count_)
|
---|
539 | rehash(bucket_count_ * bucket_count_growth_factor_);
|
---|
540 | }
|
---|
541 |
|
---|
542 | void increment_size()
|
---|
543 | {
|
---|
544 | ++size_;
|
---|
545 |
|
---|
546 | rehash_if_needed();
|
---|
547 | }
|
---|
548 |
|
---|
549 | void decrement_size()
|
---|
550 | {
|
---|
551 | --size_;
|
---|
552 | }
|
---|
553 |
|
---|
554 | private:
|
---|
555 | hash_table_bucket<value_type, size_type>* table_;
|
---|
556 | size_type bucket_count_;
|
---|
557 | size_type size_;
|
---|
558 | hasher hasher_;
|
---|
559 | key_equal key_eq_;
|
---|
560 | key_extract key_extractor_;
|
---|
561 | float max_load_factor_;
|
---|
562 |
|
---|
563 | static constexpr float bucket_count_growth_factor_{1.25};
|
---|
564 |
|
---|
565 | size_type get_bucket_idx_(const key_type& key) const
|
---|
566 | {
|
---|
567 | return hasher_(key) % bucket_count_;
|
---|
568 | }
|
---|
569 |
|
---|
570 | size_type first_filled_bucket_() const
|
---|
571 | {
|
---|
572 | size_type res{};
|
---|
573 | while (res < bucket_count_)
|
---|
574 | {
|
---|
575 | if (table_[res].head)
|
---|
576 | return res;
|
---|
577 | ++res;
|
---|
578 | }
|
---|
579 |
|
---|
580 | /**
|
---|
581 | * Note: This is used for iterators,
|
---|
582 | * so we need to return a valid index.
|
---|
583 | * But since table_[0].head is nullptr
|
---|
584 | * we know that if we return 0 the
|
---|
585 | * created iterator will test as equal
|
---|
586 | * to end().
|
---|
587 | */
|
---|
588 | return 0;
|
---|
589 | }
|
---|
590 |
|
---|
591 | friend Policy;
|
---|
592 | };
|
---|
593 | }
|
---|
594 |
|
---|
595 | #endif
|
---|