001/* 002 * Copyright (C) 2007 The Guava Authors 003 * 004 * Licensed under the Apache License, Version 2.0 (the "License"); 005 * you may not use this file except in compliance with the License. 006 * You may obtain a copy of the License at 007 * 008 * http://www.apache.org/licenses/LICENSE-2.0 009 * 010 * Unless required by applicable law or agreed to in writing, software 011 * distributed under the License is distributed on an "AS IS" BASIS, 012 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 013 * See the License for the specific language governing permissions and 014 * limitations under the License. 015 */ 016 017package com.google.common.collect; 018 019import static com.google.common.base.Preconditions.checkNotNull; 020 021import com.google.common.annotations.Beta; 022import com.google.common.annotations.GwtCompatible; 023import com.google.errorprone.annotations.CanIgnoreReturnValue; 024import com.google.errorprone.annotations.CompatibleWith; 025import java.util.Collection; 026import java.util.Collections; 027import java.util.Iterator; 028import java.util.List; 029import java.util.Set; 030import java.util.Spliterator; 031import java.util.function.Consumer; 032import java.util.function.ObjIntConsumer; 033import javax.annotation.Nullable; 034 035/** 036 * A collection that supports order-independent equality, like {@link Set}, but 037 * may have duplicate elements. A multiset is also sometimes called a 038 * <i>bag</i>. 039 * 040 * <p>Elements of a multiset that are equal to one another are referred to as 041 * <i>occurrences</i> of the same single element. The total number of 042 * occurrences of an element in a multiset is called the <i>count</i> of that 043 * element (the terms "frequency" and "multiplicity" are equivalent, but not 044 * used in this API). Since the count of an element is represented as an {@code 045 * int}, a multiset may never contain more than {@link Integer#MAX_VALUE} 046 * occurrences of any one element. 047 * 048 * <p>{@code Multiset} refines the specifications of several methods from 049 * {@code Collection}. It also defines an additional query operation, {@link 050 * #count}, which returns the count of an element. There are five new 051 * bulk-modification operations, for example {@link #add(Object, int)}, to add 052 * or remove multiple occurrences of an element at once, or to set the count of 053 * an element to a specific value. These modification operations are optional, 054 * but implementations which support the standard collection operations {@link 055 * #add(Object)} or {@link #remove(Object)} are encouraged to implement the 056 * related methods as well. Finally, two collection views are provided: {@link 057 * #elementSet} contains the distinct elements of the multiset "with duplicates 058 * collapsed", and {@link #entrySet} is similar but contains {@link Entry 059 * Multiset.Entry} instances, each providing both a distinct element and the 060 * count of that element. 061 * 062 * <p>In addition to these required methods, implementations of {@code 063 * Multiset} are expected to provide two {@code static} creation methods: 064 * {@code create()}, returning an empty multiset, and {@code 065 * create(Iterable<? extends E>)}, returning a multiset containing the 066 * given initial elements. This is simply a refinement of {@code Collection}'s 067 * constructor recommendations, reflecting the new developments of Java 5. 068 * 069 * <p>As with other collection types, the modification operations are optional, 070 * and should throw {@link UnsupportedOperationException} when they are not 071 * implemented. Most implementations should support either all add operations 072 * or none of them, all removal operations or none of them, and if and only if 073 * all of these are supported, the {@code setCount} methods as well. 074 * 075 * <p>A multiset uses {@link Object#equals} to determine whether two instances 076 * should be considered "the same," <i>unless specified otherwise</i> by the 077 * implementation. 078 * 079 * <p>Common implementations include {@link ImmutableMultiset}, {@link 080 * HashMultiset}, and {@link ConcurrentHashMultiset}. 081 * 082 * <p>If your values may be zero, negative, or outside the range of an int, you 083 * may wish to use {@link com.google.common.util.concurrent.AtomicLongMap} 084 * instead. Note, however, that unlike {@code Multiset}, {@code AtomicLongMap} 085 * does not automatically remove zeros. 086 * 087 * <p>See the Guava User Guide article on <a href= 088 * "https://github.com/google/guava/wiki/NewCollectionTypesExplained#multiset"> 089 * {@code Multiset}</a>. 090 * 091 * @author Kevin Bourrillion 092 * @since 2.0 093 */ 094@GwtCompatible 095public interface Multiset<E> extends Collection<E> { 096 // Query Operations 097 098 /** 099 * Returns the total number of all occurrences of all elements in this multiset. 100 * 101 * <p><b>Note:</b> this method does not return the number of <i>distinct elements</i> in the 102 * multiset, which is given by {@code entrySet().size()}. 103 */ 104 @Override 105 int size(); 106 107 /** 108 * Returns the number of occurrences of an element in this multiset (the <i>count</i> of the 109 * element). Note that for an {@link Object#equals}-based multiset, this gives the same result as 110 * {@link Collections#frequency} (which would presumably perform more poorly). 111 * 112 * <p><b>Note:</b> the utility method {@link Iterables#frequency} generalizes this operation; it 113 * correctly delegates to this method when dealing with a multiset, but it can also accept any 114 * other iterable type. 115 * 116 * @param element the element to count occurrences of 117 * @return the number of occurrences of the element in this multiset; possibly zero but never 118 * negative 119 */ 120 int count(@Nullable @CompatibleWith("E") Object element); 121 122 // Bulk Operations 123 124 /** 125 * Adds a number of occurrences of an element to this multiset. Note that if 126 * {@code occurrences == 1}, this method has the identical effect to {@link 127 * #add(Object)}. This method is functionally equivalent (except in the case 128 * of overflow) to the call {@code addAll(Collections.nCopies(element, 129 * occurrences))}, which would presumably perform much more poorly. 130 * 131 * @param element the element to add occurrences of; may be null only if 132 * explicitly allowed by the implementation 133 * @param occurrences the number of occurrences of the element to add. May be 134 * zero, in which case no change will be made. 135 * @return the count of the element before the operation; possibly zero 136 * @throws IllegalArgumentException if {@code occurrences} is negative, or if 137 * this operation would result in more than {@link Integer#MAX_VALUE} 138 * occurrences of the element 139 * @throws NullPointerException if {@code element} is null and this 140 * implementation does not permit null elements. Note that if {@code 141 * occurrences} is zero, the implementation may opt to return normally. 142 */ 143 @CanIgnoreReturnValue 144 int add(@Nullable E element, int occurrences); 145 146 /** 147 * Removes a number of occurrences of the specified element from this multiset. If the multiset 148 * contains fewer than this number of occurrences to begin with, all occurrences will be removed. 149 * Note that if {@code occurrences == 1}, this is functionally equivalent to the call {@code 150 * remove(element)}. 151 * 152 * @param element the element to conditionally remove occurrences of 153 * @param occurrences the number of occurrences of the element to remove. May be zero, in which 154 * case no change will be made. 155 * @return the count of the element before the operation; possibly zero 156 * @throws IllegalArgumentException if {@code occurrences} is negative 157 */ 158 @CanIgnoreReturnValue 159 int remove(@Nullable @CompatibleWith("E") Object element, int occurrences); 160 161 /** 162 * Adds or removes the necessary occurrences of an element such that the 163 * element attains the desired count. 164 * 165 * @param element the element to add or remove occurrences of; may be null 166 * only if explicitly allowed by the implementation 167 * @param count the desired count of the element in this multiset 168 * @return the count of the element before the operation; possibly zero 169 * @throws IllegalArgumentException if {@code count} is negative 170 * @throws NullPointerException if {@code element} is null and this 171 * implementation does not permit null elements. Note that if {@code 172 * count} is zero, the implementor may optionally return zero instead. 173 */ 174 @CanIgnoreReturnValue 175 int setCount(E element, int count); 176 177 /** 178 * Conditionally sets the count of an element to a new value, as described in 179 * {@link #setCount(Object, int)}, provided that the element has the expected 180 * current count. If the current count is not {@code oldCount}, no change is 181 * made. 182 * 183 * @param element the element to conditionally set the count of; may be null 184 * only if explicitly allowed by the implementation 185 * @param oldCount the expected present count of the element in this multiset 186 * @param newCount the desired count of the element in this multiset 187 * @return {@code true} if the condition for modification was met. This 188 * implies that the multiset was indeed modified, unless 189 * {@code oldCount == newCount}. 190 * @throws IllegalArgumentException if {@code oldCount} or {@code newCount} is 191 * negative 192 * @throws NullPointerException if {@code element} is null and the 193 * implementation does not permit null elements. Note that if {@code 194 * oldCount} and {@code newCount} are both zero, the implementor may 195 * optionally return {@code true} instead. 196 */ 197 @CanIgnoreReturnValue 198 boolean setCount(E element, int oldCount, int newCount); 199 200 // Views 201 202 /** 203 * Returns the set of distinct elements contained in this multiset. The 204 * element set is backed by the same data as the multiset, so any change to 205 * either is immediately reflected in the other. The order of the elements in 206 * the element set is unspecified. 207 * 208 * <p>If the element set supports any removal operations, these necessarily 209 * cause <b>all</b> occurrences of the removed element(s) to be removed from 210 * the multiset. Implementations are not expected to support the add 211 * operations, although this is possible. 212 * 213 * <p>A common use for the element set is to find the number of distinct 214 * elements in the multiset: {@code elementSet().size()}. 215 * 216 * @return a view of the set of distinct elements in this multiset 217 */ 218 Set<E> elementSet(); 219 220 /** 221 * Returns a view of the contents of this multiset, grouped into {@code 222 * Multiset.Entry} instances, each providing an element of the multiset and 223 * the count of that element. This set contains exactly one entry for each 224 * distinct element in the multiset (thus it always has the same size as the 225 * {@link #elementSet}). The order of the elements in the element set is 226 * unspecified. 227 * 228 * <p>The entry set is backed by the same data as the multiset, so any change 229 * to either is immediately reflected in the other. However, multiset changes 230 * may or may not be reflected in any {@code Entry} instances already 231 * retrieved from the entry set (this is implementation-dependent). 232 * Furthermore, implementations are not required to support modifications to 233 * the entry set at all, and the {@code Entry} instances themselves don't 234 * even have methods for modification. See the specific implementation class 235 * for more details on how its entry set handles modifications. 236 * 237 * @return a set of entries representing the data of this multiset 238 */ 239 Set<Entry<E>> entrySet(); 240 241 /** 242 * An unmodifiable element-count pair for a multiset. The {@link 243 * Multiset#entrySet} method returns a view of the multiset whose elements 244 * are of this class. A multiset implementation may return Entry instances 245 * that are either live "read-through" views to the Multiset, or immutable 246 * snapshots. Note that this type is unrelated to the similarly-named type 247 * {@code Map.Entry}. 248 * 249 * @since 2.0 250 */ 251 interface Entry<E> { 252 253 /** 254 * Returns the multiset element corresponding to this entry. Multiple calls 255 * to this method always return the same instance. 256 * 257 * @return the element corresponding to this entry 258 */ 259 E getElement(); 260 261 /** 262 * Returns the count of the associated element in the underlying multiset. 263 * This count may either be an unchanging snapshot of the count at the time 264 * the entry was retrieved, or a live view of the current count of the 265 * element in the multiset, depending on the implementation. Note that in 266 * the former case, this method can never return zero, while in the latter, 267 * it will return zero if all occurrences of the element were since removed 268 * from the multiset. 269 * 270 * @return the count of the element; never negative 271 */ 272 int getCount(); 273 274 /** 275 * {@inheritDoc} 276 * 277 * <p>Returns {@code true} if the given object is also a multiset entry and 278 * the two entries represent the same element and count. That is, two 279 * entries {@code a} and {@code b} are equal if: <pre> {@code 280 * 281 * Objects.equal(a.getElement(), b.getElement()) 282 * && a.getCount() == b.getCount()}</pre> 283 */ 284 @Override 285 // TODO(kevinb): check this wrt TreeMultiset? 286 boolean equals(Object o); 287 288 /** 289 * {@inheritDoc} 290 * 291 * <p>The hash code of a multiset entry for element {@code element} and 292 * count {@code count} is defined as: <pre> {@code 293 * 294 * ((element == null) ? 0 : element.hashCode()) ^ count}</pre> 295 */ 296 @Override 297 int hashCode(); 298 299 /** 300 * Returns the canonical string representation of this entry, defined as 301 * follows. If the count for this entry is one, this is simply the string 302 * representation of the corresponding element. Otherwise, it is the string 303 * representation of the element, followed by the three characters {@code 304 * " x "} (space, letter x, space), followed by the count. 305 */ 306 @Override 307 String toString(); 308 } 309 310 /** 311 * Runs the specified action for each distinct element in this multiset, and the number of 312 * occurrences of that element. For some {@code Multiset} implementations, this may be more 313 * efficient than iterating over the {@link #entrySet()} either explicitly or with {@code 314 * entrySet().forEach(action)}. 315 * 316 * @since 21.0 317 */ 318 @Beta 319 default void forEachEntry(ObjIntConsumer<? super E> action) { 320 checkNotNull(action); 321 entrySet().forEach(entry -> action.accept(entry.getElement(), entry.getCount())); 322 } 323 324 // Comparison and hashing 325 326 /** 327 * Compares the specified object with this multiset for equality. Returns {@code true} if the 328 * given object is also a multiset and contains equal elements with equal counts, regardless of 329 * order. 330 */ 331 @Override 332 // TODO(kevinb): caveats about equivalence-relation? 333 boolean equals(@Nullable Object object); 334 335 /** 336 * Returns the hash code for this multiset. This is defined as the sum of 337 * <pre> {@code 338 * 339 * ((element == null) ? 0 : element.hashCode()) ^ count(element)}</pre> 340 * 341 * <p>over all distinct elements in the multiset. It follows that a multiset and 342 * its entry set always have the same hash code. 343 */ 344 @Override 345 int hashCode(); 346 347 /** 348 * {@inheritDoc} 349 * 350 * <p>It is recommended, though not mandatory, that this method return the 351 * result of invoking {@link #toString} on the {@link #entrySet}, yielding a 352 * result such as {@code [a x 3, c, d x 2, e]}. 353 */ 354 @Override 355 String toString(); 356 357 // Refined Collection Methods 358 359 /** 360 * {@inheritDoc} 361 * 362 * <p>Elements that occur multiple times in the multiset will appear 363 * multiple times in this iterator, though not necessarily sequentially. 364 */ 365 @Override 366 Iterator<E> iterator(); 367 368 /** 369 * Determines whether this multiset contains the specified element. 370 * 371 * <p>This method refines {@link Collection#contains} to further specify that 372 * it <b>may not</b> throw an exception in response to {@code element} being 373 * null or of the wrong type. 374 * 375 * @param element the element to check for 376 * @return {@code true} if this multiset contains at least one occurrence of 377 * the element 378 */ 379 @Override 380 boolean contains(@Nullable Object element); 381 382 /** 383 * Returns {@code true} if this multiset contains at least one occurrence of 384 * each element in the specified collection. 385 * 386 * <p>This method refines {@link Collection#containsAll} to further specify 387 * that it <b>may not</b> throw an exception in response to any of {@code 388 * elements} being null or of the wrong type. 389 * 390 * <p><b>Note:</b> this method does not take into account the occurrence 391 * count of an element in the two collections; it may still return {@code 392 * true} even if {@code elements} contains several occurrences of an element 393 * and this multiset contains only one. This is no different than any other 394 * collection type like {@link List}, but it may be unexpected to the user of 395 * a multiset. 396 * 397 * @param elements the collection of elements to be checked for containment in 398 * this multiset 399 * @return {@code true} if this multiset contains at least one occurrence of 400 * each element contained in {@code elements} 401 * @throws NullPointerException if {@code elements} is null 402 */ 403 @Override 404 boolean containsAll(Collection<?> elements); 405 406 /** 407 * Adds a single occurrence of the specified element to this multiset. 408 * 409 * <p>This method refines {@link Collection#add}, which only <i>ensures</i> 410 * the presence of the element, to further specify that a successful call must 411 * always increment the count of the element, and the overall size of the 412 * collection, by one. 413 * 414 * <p>To both add the element and obtain the previous count of that element, 415 * use {@link #add(E, int) add}{@code (element, 1)} instead. 416 * 417 * @param element the element to add one occurrence of; may be null only if 418 * explicitly allowed by the implementation 419 * @return {@code true} always, since this call is required to modify the 420 * multiset, unlike other {@link Collection} types 421 * @throws NullPointerException if {@code element} is null and this 422 * implementation does not permit null elements 423 * @throws IllegalArgumentException if {@link Integer#MAX_VALUE} occurrences 424 * of {@code element} are already contained in this multiset 425 */ 426 @CanIgnoreReturnValue 427 @Override 428 boolean add(E element); 429 430 /** 431 * Removes a <i>single</i> occurrence of the specified element from this 432 * multiset, if present. 433 * 434 * <p>This method refines {@link Collection#remove} to further specify that it 435 * <b>may not</b> throw an exception in response to {@code element} being null 436 * or of the wrong type. 437 * 438 * <p>To both remove the element and obtain the previous count of that element, 439 * use {@link #remove(E, int) remove}{@code (element, 1)} instead. 440 * 441 * @param element the element to remove one occurrence of 442 * @return {@code true} if an occurrence was found and removed 443 */ 444 @CanIgnoreReturnValue 445 @Override 446 boolean remove(@Nullable Object element); 447 448 /** 449 * {@inheritDoc} 450 * 451 * <p><b>Note:</b> This method ignores how often any element might appear in 452 * {@code c}, and only cares whether or not an element appears at all. 453 * If you wish to remove one occurrence in this multiset for every occurrence 454 * in {@code c}, see {@link Multisets#removeOccurrences(Multiset, Multiset)}. 455 * 456 * <p>This method refines {@link Collection#removeAll} to further specify that 457 * it <b>may not</b> throw an exception in response to any of {@code elements} 458 * being null or of the wrong type. 459 */ 460 @CanIgnoreReturnValue 461 @Override 462 boolean removeAll(Collection<?> c); 463 464 /** 465 * {@inheritDoc} 466 * 467 * <p><b>Note:</b> This method ignores how often any element might appear in 468 * {@code c}, and only cares whether or not an element appears at all. 469 * If you wish to remove one occurrence in this multiset for every occurrence 470 * in {@code c}, see {@link Multisets#retainOccurrences(Multiset, Multiset)}. 471 * 472 * <p>This method refines {@link Collection#retainAll} to further specify that 473 * it <b>may not</b> throw an exception in response to any of {@code elements} 474 * being null or of the wrong type. 475 * 476 * @see Multisets#retainOccurrences(Multiset, Multiset) 477 */ 478 @CanIgnoreReturnValue 479 @Override 480 boolean retainAll(Collection<?> c); 481 482 /** 483 * {@inheritDoc} 484 * 485 * <p>Elements that occur multiple times in the multiset will be passed to the {@code Consumer} 486 * correspondingly many times, though not necessarily sequentially. 487 */ 488 @Override 489 default void forEach(Consumer<? super E> action) { 490 checkNotNull(action); 491 entrySet() 492 .forEach( 493 entry -> { 494 E elem = entry.getElement(); 495 int count = entry.getCount(); 496 for (int i = 0; i < count; i++) { 497 action.accept(elem); 498 } 499 }); 500 } 501 502 @Override 503 default Spliterator<E> spliterator() { 504 return Multisets.spliteratorImpl(this); 505 } 506}