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.GwtCompatible; 021import com.google.common.annotations.GwtIncompatible; 022import com.google.common.annotations.J2ktIncompatible; 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 java.util.stream.Stream; 031import java.util.stream.StreamSupport; 032import javax.annotation.CheckForNull; 033 034/** 035 * Extracts non-overlapping substrings from an input string, typically by recognizing appearances of 036 * a <i>separator</i> sequence. This separator can be specified as a single {@linkplain #on(char) 037 * character}, fixed {@linkplain #on(String) string}, {@linkplain #onPattern regular expression} or 038 * {@link #on(CharMatcher) CharMatcher} instance. Or, instead of using a separator at all, a 039 * splitter can extract adjacent substrings of a given {@linkplain #fixedLength fixed length}. 040 * 041 * <p>For example, this expression: 042 * 043 * <pre>{@code 044 * Splitter.on(',').split("foo,bar,qux") 045 * }</pre> 046 * 047 * ... produces an {@code Iterable} containing {@code "foo"}, {@code "bar"} and {@code "qux"}, in 048 * that order. 049 * 050 * <p>By default, {@code Splitter}'s behavior is simplistic and unassuming. The following 051 * expression: 052 * 053 * <pre>{@code 054 * Splitter.on(',').split(" foo,,, bar ,") 055 * }</pre> 056 * 057 * ... yields the substrings {@code [" foo", "", "", " bar ", ""]}. If this is not the desired 058 * behavior, use configuration methods to obtain a <i>new</i> splitter instance with modified 059 * behavior: 060 * 061 * <pre>{@code 062 * private static final Splitter MY_SPLITTER = Splitter.on(',') 063 * .trimResults() 064 * .omitEmptyStrings(); 065 * }</pre> 066 * 067 * <p>Now {@code MY_SPLITTER.split("foo,,, bar ,")} returns just {@code ["foo", "bar"]}. Note that 068 * the order in which these configuration methods are called is never significant. 069 * 070 * <p><b>Warning:</b> Splitter instances are immutable. Invoking a configuration method has no 071 * effect on the receiving instance; you must store and use the new splitter instance it returns 072 * instead. 073 * 074 * <pre>{@code 075 * // Do NOT do this 076 * Splitter splitter = Splitter.on('/'); 077 * splitter.trimResults(); // does nothing! 078 * return splitter.split("wrong / wrong / wrong"); 079 * }</pre> 080 * 081 * <p>For separator-based splitters that do not use {@code omitEmptyStrings}, an input string 082 * containing {@code n} occurrences of the separator naturally yields an iterable of size {@code n + 083 * 1}. So if the separator does not occur anywhere in the input, a single substring is returned 084 * containing the entire input. Consequently, all splitters split the empty string to {@code [""]} 085 * (note: even fixed-length splitters). 086 * 087 * <p>Splitter instances are thread-safe immutable, and are therefore safe to store as {@code static 088 * final} constants. 089 * 090 * <p>The {@link Joiner} class provides the inverse operation to splitting, but note that a 091 * round-trip between the two should be assumed to be lossy. 092 * 093 * <p>See the Guava User Guide article on <a 094 * href="https://github.com/google/guava/wiki/StringsExplained#splitter">{@code Splitter}</a>. 095 * 096 * @author Julien Silland 097 * @author Jesse Wilson 098 * @author Kevin Bourrillion 099 * @author Louis Wasserman 100 * @since 1.0 101 */ 102@GwtCompatible(emulated = true) 103@ElementTypesAreNonnullByDefault 104public final class Splitter { 105 private final CharMatcher trimmer; 106 private final boolean omitEmptyStrings; 107 private final Strategy strategy; 108 private final int limit; 109 110 private Splitter(Strategy strategy) { 111 this(strategy, false, CharMatcher.none(), Integer.MAX_VALUE); 112 } 113 114 private Splitter(Strategy strategy, boolean omitEmptyStrings, CharMatcher trimmer, int limit) { 115 this.strategy = strategy; 116 this.omitEmptyStrings = omitEmptyStrings; 117 this.trimmer = trimmer; 118 this.limit = limit; 119 } 120 121 /** 122 * Returns a splitter that uses the given single-character separator. For example, {@code 123 * Splitter.on(',').split("foo,,bar")} returns an iterable containing {@code ["foo", "", "bar"]}. 124 * 125 * @param separator the character to recognize as a separator 126 * @return a splitter, with default settings, that recognizes that separator 127 */ 128 public static Splitter on(char separator) { 129 return on(CharMatcher.is(separator)); 130 } 131 132 /** 133 * Returns a splitter that considers any single character matched by the given {@code CharMatcher} 134 * to be a separator. For example, {@code 135 * Splitter.on(CharMatcher.anyOf(";,")).split("foo,;bar,quux")} returns an iterable containing 136 * {@code ["foo", "", "bar", "quux"]}. 137 * 138 * @param separatorMatcher a {@link CharMatcher} that determines whether a character is a 139 * separator 140 * @return a splitter, with default settings, that uses this matcher 141 */ 142 public static Splitter on(final CharMatcher separatorMatcher) { 143 checkNotNull(separatorMatcher); 144 145 return new Splitter( 146 new Strategy() { 147 @Override 148 public SplittingIterator iterator(Splitter splitter, final CharSequence toSplit) { 149 return new SplittingIterator(splitter, toSplit) { 150 @Override 151 int separatorStart(int start) { 152 return separatorMatcher.indexIn(toSplit, start); 153 } 154 155 @Override 156 int separatorEnd(int separatorPosition) { 157 return separatorPosition + 1; 158 } 159 }; 160 } 161 }); 162 } 163 164 /** 165 * Returns a splitter that uses the given fixed string as a separator. For example, {@code 166 * Splitter.on(", ").split("foo, bar,baz")} returns an iterable containing {@code ["foo", 167 * "bar,baz"]}. 168 * 169 * @param separator the literal, nonempty string to recognize as a separator 170 * @return a splitter, with default settings, that recognizes that separator 171 */ 172 public static Splitter on(final String separator) { 173 checkArgument(separator.length() != 0, "The separator may not be the empty string."); 174 if (separator.length() == 1) { 175 return Splitter.on(separator.charAt(0)); 176 } 177 return new Splitter( 178 new Strategy() { 179 @Override 180 public SplittingIterator iterator(Splitter splitter, CharSequence toSplit) { 181 return new SplittingIterator(splitter, toSplit) { 182 @Override 183 public int separatorStart(int start) { 184 int separatorLength = separator.length(); 185 186 positions: 187 for (int p = start, last = toSplit.length() - separatorLength; p <= last; p++) { 188 for (int i = 0; i < separatorLength; i++) { 189 if (toSplit.charAt(i + p) != separator.charAt(i)) { 190 continue positions; 191 } 192 } 193 return p; 194 } 195 return -1; 196 } 197 198 @Override 199 public int separatorEnd(int separatorPosition) { 200 return separatorPosition + separator.length(); 201 } 202 }; 203 } 204 }); 205 } 206 207 /** 208 * Returns a splitter that considers any subsequence matching {@code pattern} to be a separator. 209 * For example, {@code Splitter.on(Pattern.compile("\r?\n")).split(entireFile)} splits a string 210 * into lines whether it uses DOS-style or UNIX-style line terminators. 211 * 212 * @param separatorPattern the pattern that determines whether a subsequence is a separator. This 213 * pattern may not match the empty string. 214 * @return a splitter, with default settings, that uses this pattern 215 * @throws IllegalArgumentException if {@code separatorPattern} matches the empty string 216 */ 217 @J2ktIncompatible 218 @GwtIncompatible // java.util.regex 219 public static Splitter on(Pattern separatorPattern) { 220 return on(new JdkPattern(separatorPattern)); 221 } 222 223 private static Splitter on(final CommonPattern separatorPattern) { 224 checkArgument( 225 !separatorPattern.matcher("").matches(), 226 "The pattern may not match the empty string: %s", 227 separatorPattern); 228 229 return new Splitter( 230 new Strategy() { 231 @Override 232 public SplittingIterator iterator(final Splitter splitter, CharSequence toSplit) { 233 final CommonMatcher matcher = separatorPattern.matcher(toSplit); 234 return new SplittingIterator(splitter, toSplit) { 235 @Override 236 public int separatorStart(int start) { 237 return matcher.find(start) ? matcher.start() : -1; 238 } 239 240 @Override 241 public int separatorEnd(int separatorPosition) { 242 return matcher.end(); 243 } 244 }; 245 } 246 }); 247 } 248 249 /** 250 * Returns a splitter that considers any subsequence matching a given pattern (regular expression) 251 * to be a separator. For example, {@code Splitter.onPattern("\r?\n").split(entireFile)} splits a 252 * string into lines whether it uses DOS-style or UNIX-style line terminators. This is equivalent 253 * to {@code Splitter.on(Pattern.compile(pattern))}. 254 * 255 * @param separatorPattern the pattern that determines whether a subsequence is a separator. This 256 * pattern may not match the empty string. 257 * @return a splitter, with default settings, that uses this pattern 258 * @throws IllegalArgumentException if {@code separatorPattern} matches the empty string or is a 259 * malformed expression 260 */ 261 @J2ktIncompatible 262 @GwtIncompatible // java.util.regex 263 public static Splitter onPattern(String separatorPattern) { 264 return on(Platform.compilePattern(separatorPattern)); 265 } 266 267 /** 268 * Returns a splitter that divides strings into pieces of the given length. For example, {@code 269 * Splitter.fixedLength(2).split("abcde")} returns an iterable containing {@code ["ab", "cd", 270 * "e"]}. The last piece can be smaller than {@code length} but will never be empty. 271 * 272 * <p><b>Note:</b> if {@link #fixedLength} is used in conjunction with {@link #limit}, the final 273 * split piece <i>may be longer than the specified fixed length</i>. This is because the splitter 274 * will <i>stop splitting when the limit is reached</i>, and just return the final piece as-is. 275 * 276 * <p><b>Exception:</b> for consistency with separator-based splitters, {@code split("")} does not 277 * yield an empty iterable, but an iterable containing {@code ""}. This is the only case in which 278 * {@code Iterables.size(split(input))} does not equal {@code IntMath.divide(input.length(), 279 * length, CEILING)}. To avoid this behavior, use {@code omitEmptyStrings}. 280 * 281 * @param length the desired length of pieces after splitting, a positive integer 282 * @return a splitter, with default settings, that can split into fixed sized pieces 283 * @throws IllegalArgumentException if {@code length} is zero or negative 284 */ 285 public static Splitter fixedLength(final int length) { 286 checkArgument(length > 0, "The length may not be less than 1"); 287 288 return new Splitter( 289 new Strategy() { 290 @Override 291 public SplittingIterator iterator(final Splitter splitter, CharSequence toSplit) { 292 return new SplittingIterator(splitter, toSplit) { 293 @Override 294 public int separatorStart(int start) { 295 int nextChunkStart = start + length; 296 return (nextChunkStart < toSplit.length() ? nextChunkStart : -1); 297 } 298 299 @Override 300 public int separatorEnd(int separatorPosition) { 301 return separatorPosition; 302 } 303 }; 304 } 305 }); 306 } 307 308 /** 309 * Returns a splitter that behaves equivalently to {@code this} splitter, but automatically omits 310 * empty strings from the results. For example, {@code 311 * Splitter.on(',').omitEmptyStrings().split(",a,,,b,c,,")} returns an iterable containing only 312 * {@code ["a", "b", "c"]}. 313 * 314 * <p>If either {@code trimResults} option is also specified when creating a splitter, that 315 * splitter always trims results first before checking for emptiness. So, for example, {@code 316 * Splitter.on(':').omitEmptyStrings().trimResults().split(": : : ")} returns an empty iterable. 317 * 318 * <p>Note that it is ordinarily not possible for {@link #split(CharSequence)} to return an empty 319 * iterable, but when using this option, it can (if the input sequence consists of nothing but 320 * separators). 321 * 322 * @return a splitter with the desired configuration 323 */ 324 public Splitter omitEmptyStrings() { 325 return new Splitter(strategy, true, trimmer, limit); 326 } 327 328 /** 329 * Returns a splitter that behaves equivalently to {@code this} splitter but stops splitting after 330 * it reaches the limit. The limit defines the maximum number of items returned by the iterator, 331 * or the maximum size of the list returned by {@link #splitToList}. 332 * 333 * <p>For example, {@code Splitter.on(',').limit(3).split("a,b,c,d")} returns an iterable 334 * containing {@code ["a", "b", "c,d"]}. When omitting empty strings, the omitted strings do not 335 * count. Hence, {@code Splitter.on(',').limit(3).omitEmptyStrings().split("a,,,b,,,c,d")} returns 336 * an iterable containing {@code ["a", "b", "c,d"}. When trim is requested, all entries are 337 * trimmed, including the last. Hence {@code Splitter.on(',').limit(3).trimResults().split(" a , b 338 * , c , d ")} results in {@code ["a", "b", "c , d"]}. 339 * 340 * @param maxItems the maximum number of items returned 341 * @return a splitter with the desired configuration 342 * @since 9.0 343 */ 344 public Splitter limit(int maxItems) { 345 checkArgument(maxItems > 0, "must be greater than zero: %s", maxItems); 346 return new Splitter(strategy, omitEmptyStrings, trimmer, maxItems); 347 } 348 349 /** 350 * Returns a splitter that behaves equivalently to {@code this} splitter, but automatically 351 * removes leading and trailing {@linkplain CharMatcher#whitespace whitespace} from each returned 352 * substring; equivalent to {@code trimResults(CharMatcher.whitespace())}. For example, {@code 353 * Splitter.on(',').trimResults().split(" a, b ,c ")} returns an iterable containing {@code ["a", 354 * "b", "c"]}. 355 * 356 * @return a splitter with the desired configuration 357 */ 358 public Splitter trimResults() { 359 return trimResults(CharMatcher.whitespace()); 360 } 361 362 /** 363 * Returns a splitter that behaves equivalently to {@code this} splitter, but removes all leading 364 * or trailing characters matching the given {@code CharMatcher} from each returned substring. For 365 * example, {@code Splitter.on(',').trimResults(CharMatcher.is('_')).split("_a ,_b_ ,c__")} 366 * returns an iterable containing {@code ["a ", "b_ ", "c"]}. 367 * 368 * @param trimmer a {@link CharMatcher} that determines whether a character should be removed from 369 * the beginning/end of a subsequence 370 * @return a splitter with the desired configuration 371 */ 372 // TODO(kevinb): throw if a trimmer was already specified! 373 public Splitter trimResults(CharMatcher trimmer) { 374 checkNotNull(trimmer); 375 return new Splitter(strategy, omitEmptyStrings, trimmer, limit); 376 } 377 378 /** 379 * Splits {@code sequence} into string components and makes them available through an {@link 380 * Iterator}, which may be lazily evaluated. If you want an eagerly computed {@link List}, use 381 * {@link #splitToList(CharSequence)}. Java 8 users may prefer {@link #splitToStream} instead. 382 * 383 * @param sequence the sequence of characters to split 384 * @return an iteration over the segments split from the parameter 385 */ 386 public Iterable<String> split(final CharSequence sequence) { 387 checkNotNull(sequence); 388 389 return new Iterable<String>() { 390 @Override 391 public Iterator<String> iterator() { 392 return splittingIterator(sequence); 393 } 394 395 @Override 396 public String toString() { 397 return Joiner.on(", ") 398 .appendTo(new StringBuilder().append('['), this) 399 .append(']') 400 .toString(); 401 } 402 }; 403 } 404 405 private Iterator<String> splittingIterator(CharSequence sequence) { 406 return strategy.iterator(this, sequence); 407 } 408 409 /** 410 * Splits {@code sequence} into string components and returns them as an immutable list. If you 411 * want an {@link Iterable} which may be lazily evaluated, use {@link #split(CharSequence)}. 412 * 413 * @param sequence the sequence of characters to split 414 * @return an immutable list of the segments split from the parameter 415 * @since 15.0 416 */ 417 public List<String> splitToList(CharSequence sequence) { 418 checkNotNull(sequence); 419 420 Iterator<String> iterator = splittingIterator(sequence); 421 List<String> result = new ArrayList<>(); 422 423 while (iterator.hasNext()) { 424 result.add(iterator.next()); 425 } 426 427 return Collections.unmodifiableList(result); 428 } 429 430 /** 431 * Splits {@code sequence} into string components and makes them available through an {@link 432 * Stream}, which may be lazily evaluated. If you want an eagerly computed {@link List}, use 433 * {@link #splitToList(CharSequence)}. 434 * 435 * @param sequence the sequence of characters to split 436 * @return a stream over the segments split from the parameter 437 * @since 28.2 438 */ 439 public Stream<String> splitToStream(CharSequence sequence) { 440 // Can't use Streams.stream() from base 441 return StreamSupport.stream(split(sequence).spliterator(), false); 442 } 443 444 /** 445 * Returns a {@code MapSplitter} which splits entries based on this splitter, and splits entries 446 * into keys and values using the specified separator. 447 * 448 * @since 10.0 449 */ 450 public MapSplitter withKeyValueSeparator(String separator) { 451 return withKeyValueSeparator(on(separator)); 452 } 453 454 /** 455 * Returns a {@code MapSplitter} which splits entries based on this splitter, and splits entries 456 * into keys and values using the specified separator. 457 * 458 * @since 14.0 459 */ 460 public MapSplitter withKeyValueSeparator(char separator) { 461 return withKeyValueSeparator(on(separator)); 462 } 463 464 /** 465 * Returns a {@code MapSplitter} which splits entries based on this splitter, and splits entries 466 * into keys and values using the specified key-value splitter. 467 * 468 * <p>Note: Any configuration option configured on this splitter, such as {@link #trimResults}, 469 * does not change the behavior of the {@code keyValueSplitter}. 470 * 471 * <p>Example: 472 * 473 * <pre>{@code 474 * String toSplit = " x -> y, z-> a "; 475 * Splitter outerSplitter = Splitter.on(',').trimResults(); 476 * MapSplitter mapSplitter = outerSplitter.withKeyValueSeparator(Splitter.on("->")); 477 * Map<String, String> result = mapSplitter.split(toSplit); 478 * assertThat(result).isEqualTo(ImmutableMap.of("x ", " y", "z", " a")); 479 * }</pre> 480 * 481 * @since 10.0 482 */ 483 public MapSplitter withKeyValueSeparator(Splitter keyValueSplitter) { 484 return new MapSplitter(this, keyValueSplitter); 485 } 486 487 /** 488 * An object that splits strings into maps as {@code Splitter} splits iterables and lists. Like 489 * {@code Splitter}, it is thread-safe and immutable. The common way to build instances is by 490 * providing an additional {@linkplain Splitter#withKeyValueSeparator key-value separator} to 491 * {@link Splitter}. 492 * 493 * @since 10.0 494 */ 495 public static final class MapSplitter { 496 private static final String INVALID_ENTRY_MESSAGE = "Chunk [%s] is not a valid entry"; 497 private final Splitter outerSplitter; 498 private final Splitter entrySplitter; 499 500 private MapSplitter(Splitter outerSplitter, Splitter entrySplitter) { 501 this.outerSplitter = outerSplitter; // only "this" is passed 502 this.entrySplitter = checkNotNull(entrySplitter); 503 } 504 505 /** 506 * Splits {@code sequence} into substrings, splits each substring into an entry, and returns an 507 * unmodifiable map with each of the entries. For example, {@code 508 * Splitter.on(';').trimResults().withKeyValueSeparator("=>").split("a=>b ; c=>b")} will return 509 * a mapping from {@code "a"} to {@code "b"} and {@code "c"} to {@code "b"}. 510 * 511 * <p>The returned map preserves the order of the entries from {@code sequence}. 512 * 513 * @throws IllegalArgumentException if the specified sequence does not split into valid map 514 * entries, or if there are duplicate keys 515 */ 516 public Map<String, String> split(CharSequence sequence) { 517 Map<String, String> map = new LinkedHashMap<>(); 518 for (String entry : outerSplitter.split(sequence)) { 519 Iterator<String> entryFields = entrySplitter.splittingIterator(entry); 520 521 checkArgument(entryFields.hasNext(), INVALID_ENTRY_MESSAGE, entry); 522 String key = entryFields.next(); 523 checkArgument(!map.containsKey(key), "Duplicate key [%s] found.", key); 524 525 checkArgument(entryFields.hasNext(), INVALID_ENTRY_MESSAGE, entry); 526 String value = entryFields.next(); 527 map.put(key, value); 528 529 checkArgument(!entryFields.hasNext(), INVALID_ENTRY_MESSAGE, entry); 530 } 531 return Collections.unmodifiableMap(map); 532 } 533 } 534 535 private interface Strategy { 536 Iterator<String> iterator(Splitter splitter, CharSequence toSplit); 537 } 538 539 private abstract static class SplittingIterator extends AbstractIterator<String> { 540 final CharSequence toSplit; 541 final CharMatcher trimmer; 542 final boolean omitEmptyStrings; 543 544 /** 545 * Returns the first index in {@code toSplit} at or after {@code start} that contains the 546 * separator. 547 */ 548 abstract int separatorStart(int start); 549 550 /** 551 * Returns the first index in {@code toSplit} after {@code separatorPosition} that does not 552 * contain a separator. This method is only invoked after a call to {@code separatorStart}. 553 */ 554 abstract int separatorEnd(int separatorPosition); 555 556 int offset = 0; 557 int limit; 558 559 protected SplittingIterator(Splitter splitter, CharSequence toSplit) { 560 this.trimmer = splitter.trimmer; 561 this.omitEmptyStrings = splitter.omitEmptyStrings; 562 this.limit = splitter.limit; 563 this.toSplit = toSplit; 564 } 565 566 @CheckForNull 567 @Override 568 protected String computeNext() { 569 /* 570 * The returned string will be from the end of the last match to the beginning of the next 571 * one. nextStart is the start position of the returned substring, while offset is the place 572 * to start looking for a separator. 573 */ 574 int nextStart = offset; 575 while (offset != -1) { 576 int start = nextStart; 577 int end; 578 579 int separatorPosition = separatorStart(offset); 580 if (separatorPosition == -1) { 581 end = toSplit.length(); 582 offset = -1; 583 } else { 584 end = separatorPosition; 585 offset = separatorEnd(separatorPosition); 586 } 587 if (offset == nextStart) { 588 /* 589 * This occurs when some pattern has an empty match, even if it doesn't match the empty 590 * string -- for example, if it requires lookahead or the like. The offset must be 591 * increased to look for separators beyond this point, without changing the start position 592 * of the next returned substring -- so nextStart stays the same. 593 */ 594 offset++; 595 if (offset > toSplit.length()) { 596 offset = -1; 597 } 598 continue; 599 } 600 601 while (start < end && trimmer.matches(toSplit.charAt(start))) { 602 start++; 603 } 604 while (end > start && trimmer.matches(toSplit.charAt(end - 1))) { 605 end--; 606 } 607 608 if (omitEmptyStrings && start == end) { 609 // Don't include the (unused) separator in next split string. 610 nextStart = offset; 611 continue; 612 } 613 614 if (limit == 1) { 615 // The limit has been reached, return the rest of the string as the 616 // final item. This is tested after empty string removal so that 617 // empty strings do not count towards the limit. 618 end = toSplit.length(); 619 offset = -1; 620 // Since we may have changed the end, we need to trim it again. 621 while (end > start && trimmer.matches(toSplit.charAt(end - 1))) { 622 end--; 623 } 624 } else { 625 limit--; 626 } 627 628 return toSplit.subSequence(start, end).toString(); 629 } 630 return endOfData(); 631 } 632 } 633}