001/* 002 * Copyright (C) 2009 The Guava Authors 003 * 004 * Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except 005 * in compliance with the License. You may obtain a copy of the License at 006 * 007 * http://www.apache.org/licenses/LICENSE-2.0 008 * 009 * Unless required by applicable law or agreed to in writing, software distributed under the License 010 * is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express 011 * or implied. See the License for the specific language governing permissions and limitations under 012 * the License. 013 */ 014 015package com.google.common.base; 016 017import static com.google.common.base.Preconditions.checkArgument; 018import static com.google.common.base.Preconditions.checkNotNull; 019 020import com.google.common.annotations.Beta; 021import com.google.common.annotations.GwtCompatible; 022import com.google.common.annotations.GwtIncompatible; 023import java.util.ArrayList; 024import java.util.Collections; 025import java.util.Iterator; 026import java.util.LinkedHashMap; 027import java.util.List; 028import java.util.Map; 029import java.util.regex.Pattern; 030import javax.annotation.CheckForNull; 031 032/** 033 * Extracts non-overlapping substrings from an input string, typically by recognizing appearances of 034 * a <i>separator</i> sequence. This separator can be specified as a single {@linkplain #on(char) 035 * character}, fixed {@linkplain #on(String) string}, {@linkplain #onPattern regular expression} or 036 * {@link #on(CharMatcher) CharMatcher} instance. Or, instead of using a separator at all, a 037 * splitter can extract adjacent substrings of a given {@linkplain #fixedLength fixed length}. 038 * 039 * <p>For example, this expression: 040 * 041 * <pre>{@code 042 * Splitter.on(',').split("foo,bar,qux") 043 * }</pre> 044 * 045 * ... produces an {@code Iterable} containing {@code "foo"}, {@code "bar"} and {@code "qux"}, in 046 * that order. 047 * 048 * <p>By default, {@code Splitter}'s behavior is simplistic and unassuming. The following 049 * expression: 050 * 051 * <pre>{@code 052 * Splitter.on(',').split(" foo,,, bar ,") 053 * }</pre> 054 * 055 * ... yields the substrings {@code [" foo", "", "", " bar ", ""]}. If this is not the desired 056 * behavior, use configuration methods to obtain a <i>new</i> splitter instance with modified 057 * behavior: 058 * 059 * <pre>{@code 060 * private static final Splitter MY_SPLITTER = Splitter.on(',') 061 * .trimResults() 062 * .omitEmptyStrings(); 063 * }</pre> 064 * 065 * <p>Now {@code MY_SPLITTER.split("foo,,, bar ,")} returns just {@code ["foo", "bar"]}. Note that 066 * the order in which these configuration methods are called is never significant. 067 * 068 * <p><b>Warning:</b> Splitter instances are immutable. Invoking a configuration method has no 069 * effect on the receiving instance; you must store and use the new splitter instance it returns 070 * instead. 071 * 072 * <pre>{@code 073 * // Do NOT do this 074 * Splitter splitter = Splitter.on('/'); 075 * splitter.trimResults(); // does nothing! 076 * return splitter.split("wrong / wrong / wrong"); 077 * }</pre> 078 * 079 * <p>For separator-based splitters that do not use {@code omitEmptyStrings}, an input string 080 * containing {@code n} occurrences of the separator naturally yields an iterable of size {@code n + 081 * 1}. So if the separator does not occur anywhere in the input, a single substring is returned 082 * containing the entire input. Consequently, all splitters split the empty string to {@code [""]} 083 * (note: even fixed-length splitters). 084 * 085 * <p>Splitter instances are thread-safe immutable, and are therefore safe to store as {@code static 086 * final} constants. 087 * 088 * <p>The {@link Joiner} class provides the inverse operation to splitting, but note that a 089 * round-trip between the two should be assumed to be lossy. 090 * 091 * <p>See the Guava User Guide article on <a 092 * href="https://github.com/google/guava/wiki/StringsExplained#splitter">{@code Splitter}</a>. 093 * 094 * @author Julien Silland 095 * @author Jesse Wilson 096 * @author Kevin Bourrillion 097 * @author Louis Wasserman 098 * @since 1.0 099 */ 100@GwtCompatible(emulated = true) 101@ElementTypesAreNonnullByDefault 102public final class Splitter { 103 private final CharMatcher trimmer; 104 private final boolean omitEmptyStrings; 105 private final Strategy strategy; 106 private final int limit; 107 108 private Splitter(Strategy strategy) { 109 this(strategy, false, CharMatcher.none(), Integer.MAX_VALUE); 110 } 111 112 private Splitter(Strategy strategy, boolean omitEmptyStrings, CharMatcher trimmer, int limit) { 113 this.strategy = strategy; 114 this.omitEmptyStrings = omitEmptyStrings; 115 this.trimmer = trimmer; 116 this.limit = limit; 117 } 118 119 /** 120 * Returns a splitter that uses the given single-character separator. For example, {@code 121 * Splitter.on(',').split("foo,,bar")} returns an iterable containing {@code ["foo", "", "bar"]}. 122 * 123 * @param separator the character to recognize as a separator 124 * @return a splitter, with default settings, that recognizes that separator 125 */ 126 public static Splitter on(char separator) { 127 return on(CharMatcher.is(separator)); 128 } 129 130 /** 131 * Returns a splitter that considers any single character matched by the given {@code CharMatcher} 132 * to be a separator. For example, {@code 133 * Splitter.on(CharMatcher.anyOf(";,")).split("foo,;bar,quux")} returns an iterable containing 134 * {@code ["foo", "", "bar", "quux"]}. 135 * 136 * @param separatorMatcher a {@link CharMatcher} that determines whether a character is a 137 * separator 138 * @return a splitter, with default settings, that uses this matcher 139 */ 140 public static Splitter on(final CharMatcher separatorMatcher) { 141 checkNotNull(separatorMatcher); 142 143 return new Splitter( 144 new Strategy() { 145 @Override 146 public SplittingIterator iterator(Splitter splitter, final CharSequence toSplit) { 147 return new SplittingIterator(splitter, toSplit) { 148 @Override 149 int separatorStart(int start) { 150 return separatorMatcher.indexIn(toSplit, start); 151 } 152 153 @Override 154 int separatorEnd(int separatorPosition) { 155 return separatorPosition + 1; 156 } 157 }; 158 } 159 }); 160 } 161 162 /** 163 * Returns a splitter that uses the given fixed string as a separator. For example, {@code 164 * Splitter.on(", ").split("foo, bar,baz")} returns an iterable containing {@code ["foo", 165 * "bar,baz"]}. 166 * 167 * @param separator the literal, nonempty string to recognize as a separator 168 * @return a splitter, with default settings, that recognizes that separator 169 */ 170 public static Splitter on(final String separator) { 171 checkArgument(separator.length() != 0, "The separator may not be the empty string."); 172 if (separator.length() == 1) { 173 return Splitter.on(separator.charAt(0)); 174 } 175 return new Splitter( 176 new Strategy() { 177 @Override 178 public SplittingIterator iterator(Splitter splitter, CharSequence toSplit) { 179 return new SplittingIterator(splitter, toSplit) { 180 @Override 181 public int separatorStart(int start) { 182 int separatorLength = separator.length(); 183 184 positions: 185 for (int p = start, last = toSplit.length() - separatorLength; p <= last; p++) { 186 for (int i = 0; i < separatorLength; i++) { 187 if (toSplit.charAt(i + p) != separator.charAt(i)) { 188 continue positions; 189 } 190 } 191 return p; 192 } 193 return -1; 194 } 195 196 @Override 197 public int separatorEnd(int separatorPosition) { 198 return separatorPosition + separator.length(); 199 } 200 }; 201 } 202 }); 203 } 204 205 /** 206 * Returns a splitter that considers any subsequence matching {@code pattern} to be a separator. 207 * For example, {@code Splitter.on(Pattern.compile("\r?\n")).split(entireFile)} splits a string 208 * into lines whether it uses DOS-style or UNIX-style line terminators. 209 * 210 * @param separatorPattern the pattern that determines whether a subsequence is a separator. This 211 * pattern may not match the empty string. 212 * @return a splitter, with default settings, that uses this pattern 213 * @throws IllegalArgumentException if {@code separatorPattern} matches the empty string 214 */ 215 @GwtIncompatible // java.util.regex 216 public static Splitter on(Pattern separatorPattern) { 217 return on(new JdkPattern(separatorPattern)); 218 } 219 220 private static Splitter on(final CommonPattern separatorPattern) { 221 checkArgument( 222 !separatorPattern.matcher("").matches(), 223 "The pattern may not match the empty string: %s", 224 separatorPattern); 225 226 return new Splitter( 227 new Strategy() { 228 @Override 229 public SplittingIterator iterator(final Splitter splitter, CharSequence toSplit) { 230 final CommonMatcher matcher = separatorPattern.matcher(toSplit); 231 return new SplittingIterator(splitter, toSplit) { 232 @Override 233 public int separatorStart(int start) { 234 return matcher.find(start) ? matcher.start() : -1; 235 } 236 237 @Override 238 public int separatorEnd(int separatorPosition) { 239 return matcher.end(); 240 } 241 }; 242 } 243 }); 244 } 245 246 /** 247 * Returns a splitter that considers any subsequence matching a given pattern (regular expression) 248 * to be a separator. For example, {@code Splitter.onPattern("\r?\n").split(entireFile)} splits a 249 * string into lines whether it uses DOS-style or UNIX-style line terminators. This is equivalent 250 * to {@code Splitter.on(Pattern.compile(pattern))}. 251 * 252 * @param separatorPattern the pattern that determines whether a subsequence is a separator. This 253 * pattern may not match the empty string. 254 * @return a splitter, with default settings, that uses this pattern 255 * @throws IllegalArgumentException if {@code separatorPattern} matches the empty string or is a 256 * malformed expression 257 */ 258 @GwtIncompatible // java.util.regex 259 public static Splitter onPattern(String separatorPattern) { 260 return on(Platform.compilePattern(separatorPattern)); 261 } 262 263 /** 264 * Returns a splitter that divides strings into pieces of the given length. For example, {@code 265 * Splitter.fixedLength(2).split("abcde")} returns an iterable containing {@code ["ab", "cd", 266 * "e"]}. The last piece can be smaller than {@code length} but will never be empty. 267 * 268 * <p><b>Note:</b> if {@link #fixedLength} is used in conjunction with {@link #limit}, the final 269 * split piece <i>may be longer than the specified fixed length</i>. This is because the splitter 270 * will <i>stop splitting when the limit is reached</i>, and just return the final piece as-is. 271 * 272 * <p><b>Exception:</b> for consistency with separator-based splitters, {@code split("")} does not 273 * yield an empty iterable, but an iterable containing {@code ""}. This is the only case in which 274 * {@code Iterables.size(split(input))} does not equal {@code IntMath.divide(input.length(), 275 * length, CEILING)}. To avoid this behavior, use {@code omitEmptyStrings}. 276 * 277 * @param length the desired length of pieces after splitting, a positive integer 278 * @return a splitter, with default settings, that can split into fixed sized pieces 279 * @throws IllegalArgumentException if {@code length} is zero or negative 280 */ 281 public static Splitter fixedLength(final int length) { 282 checkArgument(length > 0, "The length may not be less than 1"); 283 284 return new Splitter( 285 new Strategy() { 286 @Override 287 public SplittingIterator iterator(final Splitter splitter, CharSequence toSplit) { 288 return new SplittingIterator(splitter, toSplit) { 289 @Override 290 public int separatorStart(int start) { 291 int nextChunkStart = start + length; 292 return (nextChunkStart < toSplit.length() ? nextChunkStart : -1); 293 } 294 295 @Override 296 public int separatorEnd(int separatorPosition) { 297 return separatorPosition; 298 } 299 }; 300 } 301 }); 302 } 303 304 /** 305 * Returns a splitter that behaves equivalently to {@code this} splitter, but automatically omits 306 * empty strings from the results. For example, {@code 307 * Splitter.on(',').omitEmptyStrings().split(",a,,,b,c,,")} returns an iterable containing only 308 * {@code ["a", "b", "c"]}. 309 * 310 * <p>If either {@code trimResults} option is also specified when creating a splitter, that 311 * splitter always trims results first before checking for emptiness. So, for example, {@code 312 * Splitter.on(':').omitEmptyStrings().trimResults().split(": : : ")} returns an empty iterable. 313 * 314 * <p>Note that it is ordinarily not possible for {@link #split(CharSequence)} to return an empty 315 * iterable, but when using this option, it can (if the input sequence consists of nothing but 316 * separators). 317 * 318 * @return a splitter with the desired configuration 319 */ 320 public Splitter omitEmptyStrings() { 321 return new Splitter(strategy, true, trimmer, limit); 322 } 323 324 /** 325 * Returns a splitter that behaves equivalently to {@code this} splitter but stops splitting after 326 * it reaches the limit. The limit defines the maximum number of items returned by the iterator, 327 * or the maximum size of the list returned by {@link #splitToList}. 328 * 329 * <p>For example, {@code Splitter.on(',').limit(3).split("a,b,c,d")} returns an iterable 330 * containing {@code ["a", "b", "c,d"]}. When omitting empty strings, the omitted strings do not 331 * count. Hence, {@code Splitter.on(',').limit(3).omitEmptyStrings().split("a,,,b,,,c,d")} returns 332 * an iterable containing {@code ["a", "b", "c,d"}. When trim is requested, all entries are 333 * trimmed, including the last. Hence {@code Splitter.on(',').limit(3).trimResults().split(" a , b 334 * , c , d ")} results in {@code ["a", "b", "c , d"]}. 335 * 336 * @param maxItems the maximum number of items returned 337 * @return a splitter with the desired configuration 338 * @since 9.0 339 */ 340 public Splitter limit(int maxItems) { 341 checkArgument(maxItems > 0, "must be greater than zero: %s", maxItems); 342 return new Splitter(strategy, omitEmptyStrings, trimmer, maxItems); 343 } 344 345 /** 346 * Returns a splitter that behaves equivalently to {@code this} splitter, but automatically 347 * removes leading and trailing {@linkplain CharMatcher#whitespace whitespace} from each returned 348 * substring; equivalent to {@code trimResults(CharMatcher.whitespace())}. For example, {@code 349 * Splitter.on(',').trimResults().split(" a, b ,c ")} returns an iterable containing {@code ["a", 350 * "b", "c"]}. 351 * 352 * @return a splitter with the desired configuration 353 */ 354 public Splitter trimResults() { 355 return trimResults(CharMatcher.whitespace()); 356 } 357 358 /** 359 * Returns a splitter that behaves equivalently to {@code this} splitter, but removes all leading 360 * or trailing characters matching the given {@code CharMatcher} from each returned substring. For 361 * example, {@code Splitter.on(',').trimResults(CharMatcher.is('_')).split("_a ,_b_ ,c__")} 362 * returns an iterable containing {@code ["a ", "b_ ", "c"]}. 363 * 364 * @param trimmer a {@link CharMatcher} that determines whether a character should be removed from 365 * the beginning/end of a subsequence 366 * @return a splitter with the desired configuration 367 */ 368 // TODO(kevinb): throw if a trimmer was already specified! 369 public Splitter trimResults(CharMatcher trimmer) { 370 checkNotNull(trimmer); 371 return new Splitter(strategy, omitEmptyStrings, trimmer, limit); 372 } 373 374 /** 375 * Splits {@code sequence} into string components and makes them available through an {@link 376 * Iterator}, which may be lazily evaluated. If you want an eagerly computed {@link List}, use 377 * {@link #splitToList(CharSequence)}. 378 * 379 * @param sequence the sequence of characters to split 380 * @return an iteration over the segments split from the parameter 381 */ 382 public Iterable<String> split(final CharSequence sequence) { 383 checkNotNull(sequence); 384 385 return new Iterable<String>() { 386 @Override 387 public Iterator<String> iterator() { 388 return splittingIterator(sequence); 389 } 390 391 @Override 392 public String toString() { 393 return Joiner.on(", ") 394 .appendTo(new StringBuilder().append('['), this) 395 .append(']') 396 .toString(); 397 } 398 }; 399 } 400 401 private Iterator<String> splittingIterator(CharSequence sequence) { 402 return strategy.iterator(this, sequence); 403 } 404 405 /** 406 * Splits {@code sequence} into string components and returns them as an immutable list. If you 407 * want an {@link Iterable} which may be lazily evaluated, use {@link #split(CharSequence)}. 408 * 409 * @param sequence the sequence of characters to split 410 * @return an immutable list of the segments split from the parameter 411 * @since 15.0 412 */ 413 public List<String> splitToList(CharSequence sequence) { 414 checkNotNull(sequence); 415 416 Iterator<String> iterator = splittingIterator(sequence); 417 List<String> result = new ArrayList<>(); 418 419 while (iterator.hasNext()) { 420 result.add(iterator.next()); 421 } 422 423 return Collections.unmodifiableList(result); 424 } 425 426 /** 427 * Returns a {@code MapSplitter} which splits entries based on this splitter, and splits entries 428 * into keys and values using the specified separator. 429 * 430 * @since 10.0 431 */ 432 @Beta 433 public MapSplitter withKeyValueSeparator(String separator) { 434 return withKeyValueSeparator(on(separator)); 435 } 436 437 /** 438 * Returns a {@code MapSplitter} which splits entries based on this splitter, and splits entries 439 * into keys and values using the specified separator. 440 * 441 * @since 14.0 442 */ 443 @Beta 444 public MapSplitter withKeyValueSeparator(char separator) { 445 return withKeyValueSeparator(on(separator)); 446 } 447 448 /** 449 * Returns a {@code MapSplitter} which splits entries based on this splitter, and splits entries 450 * into keys and values using the specified key-value splitter. 451 * 452 * <p>Note: Any configuration option configured on this splitter, such as {@link #trimResults}, 453 * does not change the behavior of the {@code keyValueSplitter}. 454 * 455 * <p>Example: 456 * 457 * <pre>{@code 458 * String toSplit = " x -> y, z-> a "; 459 * Splitter outerSplitter = Splitter.on(',').trimResults(); 460 * MapSplitter mapSplitter = outerSplitter.withKeyValueSeparator(Splitter.on("->")); 461 * Map<String, String> result = mapSplitter.split(toSplit); 462 * assertThat(result).isEqualTo(ImmutableMap.of("x ", " y", "z", " a")); 463 * }</pre> 464 * 465 * @since 10.0 466 */ 467 @Beta 468 public MapSplitter withKeyValueSeparator(Splitter keyValueSplitter) { 469 return new MapSplitter(this, keyValueSplitter); 470 } 471 472 /** 473 * An object that splits strings into maps as {@code Splitter} splits iterables and lists. Like 474 * {@code Splitter}, it is thread-safe and immutable. The common way to build instances is by 475 * providing an additional {@linkplain Splitter#withKeyValueSeparator key-value separator} to 476 * {@link Splitter}. 477 * 478 * @since 10.0 479 */ 480 @Beta 481 public static final class MapSplitter { 482 private static final String INVALID_ENTRY_MESSAGE = "Chunk [%s] is not a valid entry"; 483 private final Splitter outerSplitter; 484 private final Splitter entrySplitter; 485 486 private MapSplitter(Splitter outerSplitter, Splitter entrySplitter) { 487 this.outerSplitter = outerSplitter; // only "this" is passed 488 this.entrySplitter = checkNotNull(entrySplitter); 489 } 490 491 /** 492 * Splits {@code sequence} into substrings, splits each substring into an entry, and returns an 493 * unmodifiable map with each of the entries. For example, {@code 494 * Splitter.on(';').trimResults().withKeyValueSeparator("=>").split("a=>b ; c=>b")} will return 495 * a mapping from {@code "a"} to {@code "b"} and {@code "c"} to {@code "b"}. 496 * 497 * <p>The returned map preserves the order of the entries from {@code sequence}. 498 * 499 * @throws IllegalArgumentException if the specified sequence does not split into valid map 500 * entries, or if there are duplicate keys 501 */ 502 public Map<String, String> split(CharSequence sequence) { 503 Map<String, String> map = new LinkedHashMap<>(); 504 for (String entry : outerSplitter.split(sequence)) { 505 Iterator<String> entryFields = entrySplitter.splittingIterator(entry); 506 507 checkArgument(entryFields.hasNext(), INVALID_ENTRY_MESSAGE, entry); 508 String key = entryFields.next(); 509 checkArgument(!map.containsKey(key), "Duplicate key [%s] found.", key); 510 511 checkArgument(entryFields.hasNext(), INVALID_ENTRY_MESSAGE, entry); 512 String value = entryFields.next(); 513 map.put(key, value); 514 515 checkArgument(!entryFields.hasNext(), INVALID_ENTRY_MESSAGE, entry); 516 } 517 return Collections.unmodifiableMap(map); 518 } 519 } 520 521 private interface Strategy { 522 Iterator<String> iterator(Splitter splitter, CharSequence toSplit); 523 } 524 525 private abstract static class SplittingIterator extends AbstractIterator<String> { 526 final CharSequence toSplit; 527 final CharMatcher trimmer; 528 final boolean omitEmptyStrings; 529 530 /** 531 * Returns the first index in {@code toSplit} at or after {@code start} that contains the 532 * separator. 533 */ 534 abstract int separatorStart(int start); 535 536 /** 537 * Returns the first index in {@code toSplit} after {@code separatorPosition} that does not 538 * contain a separator. This method is only invoked after a call to {@code separatorStart}. 539 */ 540 abstract int separatorEnd(int separatorPosition); 541 542 int offset = 0; 543 int limit; 544 545 protected SplittingIterator(Splitter splitter, CharSequence toSplit) { 546 this.trimmer = splitter.trimmer; 547 this.omitEmptyStrings = splitter.omitEmptyStrings; 548 this.limit = splitter.limit; 549 this.toSplit = toSplit; 550 } 551 552 @CheckForNull 553 @Override 554 protected String computeNext() { 555 /* 556 * The returned string will be from the end of the last match to the beginning of the next 557 * one. nextStart is the start position of the returned substring, while offset is the place 558 * to start looking for a separator. 559 */ 560 int nextStart = offset; 561 while (offset != -1) { 562 int start = nextStart; 563 int end; 564 565 int separatorPosition = separatorStart(offset); 566 if (separatorPosition == -1) { 567 end = toSplit.length(); 568 offset = -1; 569 } else { 570 end = separatorPosition; 571 offset = separatorEnd(separatorPosition); 572 } 573 if (offset == nextStart) { 574 /* 575 * This occurs when some pattern has an empty match, even if it doesn't match the empty 576 * string -- for example, if it requires lookahead or the like. The offset must be 577 * increased to look for separators beyond this point, without changing the start position 578 * of the next returned substring -- so nextStart stays the same. 579 */ 580 offset++; 581 if (offset > toSplit.length()) { 582 offset = -1; 583 } 584 continue; 585 } 586 587 while (start < end && trimmer.matches(toSplit.charAt(start))) { 588 start++; 589 } 590 while (end > start && trimmer.matches(toSplit.charAt(end - 1))) { 591 end--; 592 } 593 594 if (omitEmptyStrings && start == end) { 595 // Don't include the (unused) separator in next split string. 596 nextStart = offset; 597 continue; 598 } 599 600 if (limit == 1) { 601 // The limit has been reached, return the rest of the string as the 602 // final item. This is tested after empty string removal so that 603 // empty strings do not count towards the limit. 604 end = toSplit.length(); 605 offset = -1; 606 // Since we may have changed the end, we need to trim it again. 607 while (end > start && trimmer.matches(toSplit.charAt(end - 1))) { 608 end--; 609 } 610 } else { 611 limit--; 612 } 613 614 return toSplit.subSequence(start, end).toString(); 615 } 616 return endOfData(); 617 } 618 } 619}