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