001/* 002 * Licensed to the Apache Software Foundation (ASF) under one 003 * or more contributor license agreements. See the NOTICE file 004 * distributed with this work for additional information 005 * regarding copyright ownership. The ASF licenses this file 006 * to you under the Apache License, Version 2.0 (the 007 * "License"); you may not use this file except in compliance 008 * with the License. You may obtain a copy of the License at 009 * 010 * http://www.apache.org/licenses/LICENSE-2.0 011 * 012 * Unless required by applicable law or agreed to in writing, software 013 * distributed under the License is distributed on an "AS IS" BASIS, 014 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 015 * See the License for the specific language governing permissions and 016 * limitations under the License. 017 */ 018package org.apache.hadoop.hbase.client; 019 020import java.io.IOException; 021import java.util.ArrayList; 022import java.util.HashMap; 023import java.util.List; 024import java.util.Map; 025import java.util.NavigableSet; 026import java.util.TreeMap; 027import java.util.TreeSet; 028import java.util.stream.Collectors; 029import org.apache.hadoop.hbase.HConstants; 030import org.apache.hadoop.hbase.client.metrics.ScanMetrics; 031import org.apache.hadoop.hbase.filter.Filter; 032import org.apache.hadoop.hbase.filter.IncompatibleFilterException; 033import org.apache.hadoop.hbase.io.TimeRange; 034import org.apache.hadoop.hbase.security.access.Permission; 035import org.apache.hadoop.hbase.security.visibility.Authorizations; 036import org.apache.hadoop.hbase.util.Bytes; 037import org.apache.yetus.audience.InterfaceAudience; 038import org.slf4j.Logger; 039import org.slf4j.LoggerFactory; 040 041import org.apache.hadoop.hbase.shaded.protobuf.ProtobufUtil; 042 043/** 044 * Used to perform Scan operations. 045 * <p> 046 * All operations are identical to {@link Get} with the exception of instantiation. Rather than 047 * specifying a single row, an optional startRow and stopRow may be defined. If rows are not 048 * specified, the Scanner will iterate over all rows. 049 * <p> 050 * To get all columns from all rows of a Table, create an instance with no constraints; use the 051 * {@link #Scan()} constructor. To constrain the scan to specific column families, call 052 * {@link #addFamily(byte[]) addFamily} for each family to retrieve on your Scan instance. 053 * <p> 054 * To get specific columns, call {@link #addColumn(byte[], byte[]) addColumn} for each column to 055 * retrieve. 056 * <p> 057 * To only retrieve columns within a specific range of version timestamps, call 058 * {@link #setTimeRange(long, long) setTimeRange}. 059 * <p> 060 * To only retrieve columns with a specific timestamp, call {@link #setTimestamp(long) setTimestamp} 061 * . 062 * <p> 063 * To limit the number of versions of each column to be returned, call {@link #setMaxVersions(int) 064 * setMaxVersions}. 065 * <p> 066 * To limit the maximum number of values returned for each call to next(), call 067 * {@link #setBatch(int) setBatch}. 068 * <p> 069 * To add a filter, call {@link #setFilter(org.apache.hadoop.hbase.filter.Filter) setFilter}. 070 * <p> 071 * For small scan, it is deprecated in 2.0.0. Now we have a {@link #setLimit(int)} method in Scan 072 * object which is used to tell RS how many rows we want. If the rows return reaches the limit, the 073 * RS will close the RegionScanner automatically. And we will also fetch data when openScanner in 074 * the new implementation, this means we can also finish a scan operation in one rpc call. And we 075 * have also introduced a {@link #setReadType(ReadType)} method. You can use this method to tell RS 076 * to use pread explicitly. 077 * <p> 078 * Expert: To explicitly disable server-side block caching for this scan, execute 079 * {@link #setCacheBlocks(boolean)}. 080 * <p> 081 * <em>Note:</em> Usage alters Scan instances. Internally, attributes are updated as the Scan runs 082 * and if enabled, metrics accumulate in the Scan instance. Be aware this is the case when you go to 083 * clone a Scan instance or if you go to reuse a created Scan instance; safer is create a Scan 084 * instance per usage. 085 */ 086@InterfaceAudience.Public 087public class Scan extends Query { 088 private static final Logger LOG = LoggerFactory.getLogger(Scan.class); 089 090 private static final String RAW_ATTR = "_raw_"; 091 092 private byte[] startRow = HConstants.EMPTY_START_ROW; 093 private boolean includeStartRow = true; 094 private byte[] stopRow = HConstants.EMPTY_END_ROW; 095 private boolean includeStopRow = false; 096 private int maxVersions = 1; 097 private int batch = -1; 098 099 /** 100 * Partial {@link Result}s are {@link Result}s must be combined to form a complete {@link Result}. 101 * The {@link Result}s had to be returned in fragments (i.e. as partials) because the size of the 102 * cells in the row exceeded max result size on the server. Typically partial results will be 103 * combined client side into complete results before being delivered to the caller. However, if 104 * this flag is set, the caller is indicating that they do not mind seeing partial results (i.e. 105 * they understand that the results returned from the Scanner may only represent part of a 106 * particular row). In such a case, any attempt to combine the partials into a complete result on 107 * the client side will be skipped, and the caller will be able to see the exact results returned 108 * from the server. 109 */ 110 private boolean allowPartialResults = false; 111 112 private int storeLimit = -1; 113 private int storeOffset = 0; 114 115 /** 116 * @deprecated since 1.0.0. Use {@link #setScanMetricsEnabled(boolean)} 117 */ 118 // Make private or remove. 119 @Deprecated 120 static public final String SCAN_ATTRIBUTES_METRICS_ENABLE = "scan.attributes.metrics.enable"; 121 122 /** 123 * Use {@link #getScanMetrics()} 124 */ 125 // Make this private or remove. 126 @Deprecated 127 static public final String SCAN_ATTRIBUTES_METRICS_DATA = "scan.attributes.metrics.data"; 128 129 // If an application wants to use multiple scans over different tables each scan must 130 // define this attribute with the appropriate table name by calling 131 // scan.setAttribute(Scan.SCAN_ATTRIBUTES_TABLE_NAME, Bytes.toBytes(tableName)) 132 static public final String SCAN_ATTRIBUTES_TABLE_NAME = "scan.attributes.table.name"; 133 134 /** 135 * -1 means no caching specified and the value of {@link HConstants#HBASE_CLIENT_SCANNER_CACHING} 136 * (default to {@link HConstants#DEFAULT_HBASE_CLIENT_SCANNER_CACHING}) will be used 137 */ 138 private int caching = -1; 139 private long maxResultSize = -1; 140 private boolean cacheBlocks = true; 141 private boolean reversed = false; 142 private TimeRange tr = TimeRange.allTime(); 143 private Map<byte[], NavigableSet<byte[]>> familyMap = 144 new TreeMap<byte[], NavigableSet<byte[]>>(Bytes.BYTES_COMPARATOR); 145 private Boolean asyncPrefetch = null; 146 147 /** 148 * Parameter name for client scanner sync/async prefetch toggle. When using async scanner, 149 * prefetching data from the server is done at the background. The parameter currently won't have 150 * any effect in the case that the user has set Scan#setSmall or Scan#setReversed 151 */ 152 public static final String HBASE_CLIENT_SCANNER_ASYNC_PREFETCH = 153 "hbase.client.scanner.async.prefetch"; 154 155 /** 156 * Default value of {@link #HBASE_CLIENT_SCANNER_ASYNC_PREFETCH}. 157 */ 158 public static final boolean DEFAULT_HBASE_CLIENT_SCANNER_ASYNC_PREFETCH = false; 159 160 /** 161 * Set it true for small scan to get better performance Small scan should use pread and big scan 162 * can use seek + read seek + read is fast but can cause two problem (1) resource contention (2) 163 * cause too much network io [89-fb] Using pread for non-compaction read request 164 * https://issues.apache.org/jira/browse/HBASE-7266 On the other hand, if setting it true, we 165 * would do openScanner,next,closeScanner in one RPC call. It means the better performance for 166 * small scan. [HBASE-9488]. Generally, if the scan range is within one data block(64KB), it could 167 * be considered as a small scan. 168 */ 169 private boolean small = false; 170 171 /** 172 * The mvcc read point to use when open a scanner. Remember to clear it after switching regions as 173 * the mvcc is only valid within region scope. 174 */ 175 private long mvccReadPoint = -1L; 176 177 /** 178 * The number of rows we want for this scan. We will terminate the scan if the number of return 179 * rows reaches this value. 180 */ 181 private int limit = -1; 182 183 /** 184 * Control whether to use pread at server side. 185 */ 186 private ReadType readType = ReadType.DEFAULT; 187 188 private boolean needCursorResult = false; 189 190 /** 191 * Create a Scan operation across all rows. 192 */ 193 public Scan() { 194 } 195 196 /** 197 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use 198 * {@code new Scan().withStartRow(startRow).setFilter(filter)} instead. 199 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 200 */ 201 @Deprecated 202 public Scan(byte[] startRow, Filter filter) { 203 this(startRow); 204 this.filter = filter; 205 } 206 207 /** 208 * Create a Scan operation starting at the specified row. 209 * <p> 210 * If the specified row does not exist, the Scanner will start from the next closest row after the 211 * specified row. 212 * @param startRow row to start scanner at or after 213 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use 214 * {@code new Scan().withStartRow(startRow)} instead. 215 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 216 */ 217 @Deprecated 218 public Scan(byte[] startRow) { 219 setStartRow(startRow); 220 } 221 222 /** 223 * Create a Scan operation for the range of rows specified. 224 * @param startRow row to start scanner at or after (inclusive) 225 * @param stopRow row to stop scanner before (exclusive) 226 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use 227 * {@code new Scan().withStartRow(startRow).withStopRow(stopRow)} instead. 228 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 229 */ 230 @Deprecated 231 public Scan(byte[] startRow, byte[] stopRow) { 232 setStartRow(startRow); 233 setStopRow(stopRow); 234 } 235 236 /** 237 * Creates a new instance of this class while copying all values. 238 * @param scan The scan instance to copy from. 239 * @throws IOException When copying the values fails. 240 */ 241 public Scan(Scan scan) throws IOException { 242 startRow = scan.getStartRow(); 243 includeStartRow = scan.includeStartRow(); 244 stopRow = scan.getStopRow(); 245 includeStopRow = scan.includeStopRow(); 246 maxVersions = scan.getMaxVersions(); 247 batch = scan.getBatch(); 248 storeLimit = scan.getMaxResultsPerColumnFamily(); 249 storeOffset = scan.getRowOffsetPerColumnFamily(); 250 caching = scan.getCaching(); 251 maxResultSize = scan.getMaxResultSize(); 252 cacheBlocks = scan.getCacheBlocks(); 253 filter = scan.getFilter(); // clone? 254 loadColumnFamiliesOnDemand = scan.getLoadColumnFamiliesOnDemandValue(); 255 consistency = scan.getConsistency(); 256 this.setIsolationLevel(scan.getIsolationLevel()); 257 reversed = scan.isReversed(); 258 asyncPrefetch = scan.isAsyncPrefetch(); 259 small = scan.isSmall(); 260 allowPartialResults = scan.getAllowPartialResults(); 261 tr = scan.getTimeRange(); // TimeRange is immutable 262 Map<byte[], NavigableSet<byte[]>> fams = scan.getFamilyMap(); 263 for (Map.Entry<byte[], NavigableSet<byte[]>> entry : fams.entrySet()) { 264 byte[] fam = entry.getKey(); 265 NavigableSet<byte[]> cols = entry.getValue(); 266 if (cols != null && cols.size() > 0) { 267 for (byte[] col : cols) { 268 addColumn(fam, col); 269 } 270 } else { 271 addFamily(fam); 272 } 273 } 274 for (Map.Entry<String, byte[]> attr : scan.getAttributesMap().entrySet()) { 275 setAttribute(attr.getKey(), attr.getValue()); 276 } 277 for (Map.Entry<byte[], TimeRange> entry : scan.getColumnFamilyTimeRange().entrySet()) { 278 TimeRange tr = entry.getValue(); 279 setColumnFamilyTimeRange(entry.getKey(), tr.getMin(), tr.getMax()); 280 } 281 this.mvccReadPoint = scan.getMvccReadPoint(); 282 this.limit = scan.getLimit(); 283 this.needCursorResult = scan.isNeedCursorResult(); 284 setPriority(scan.getPriority()); 285 readType = scan.getReadType(); 286 super.setReplicaId(scan.getReplicaId()); 287 } 288 289 /** 290 * Builds a scan object with the same specs as get. 291 * @param get get to model scan after 292 */ 293 public Scan(Get get) { 294 this.startRow = get.getRow(); 295 this.includeStartRow = true; 296 this.stopRow = get.getRow(); 297 this.includeStopRow = true; 298 this.filter = get.getFilter(); 299 this.cacheBlocks = get.getCacheBlocks(); 300 this.maxVersions = get.getMaxVersions(); 301 this.storeLimit = get.getMaxResultsPerColumnFamily(); 302 this.storeOffset = get.getRowOffsetPerColumnFamily(); 303 this.tr = get.getTimeRange(); 304 this.familyMap = get.getFamilyMap(); 305 this.asyncPrefetch = false; 306 this.consistency = get.getConsistency(); 307 this.setIsolationLevel(get.getIsolationLevel()); 308 this.loadColumnFamiliesOnDemand = get.getLoadColumnFamiliesOnDemandValue(); 309 for (Map.Entry<String, byte[]> attr : get.getAttributesMap().entrySet()) { 310 setAttribute(attr.getKey(), attr.getValue()); 311 } 312 for (Map.Entry<byte[], TimeRange> entry : get.getColumnFamilyTimeRange().entrySet()) { 313 TimeRange tr = entry.getValue(); 314 setColumnFamilyTimeRange(entry.getKey(), tr.getMin(), tr.getMax()); 315 } 316 this.mvccReadPoint = -1L; 317 setPriority(get.getPriority()); 318 super.setReplicaId(get.getReplicaId()); 319 } 320 321 public boolean isGetScan() { 322 return includeStartRow && includeStopRow 323 && ClientUtil.areScanStartRowAndStopRowEqual(this.startRow, this.stopRow); 324 } 325 326 /** 327 * Get all columns from the specified family. 328 * <p> 329 * Overrides previous calls to addColumn for this family. 330 * @param family family name 331 */ 332 public Scan addFamily(byte[] family) { 333 familyMap.remove(family); 334 familyMap.put(family, null); 335 return this; 336 } 337 338 /** 339 * Get the column from the specified family with the specified qualifier. 340 * <p> 341 * Overrides previous calls to addFamily for this family. 342 * @param family family name 343 * @param qualifier column qualifier 344 */ 345 public Scan addColumn(byte[] family, byte[] qualifier) { 346 NavigableSet<byte[]> set = familyMap.get(family); 347 if (set == null) { 348 set = new TreeSet<>(Bytes.BYTES_COMPARATOR); 349 familyMap.put(family, set); 350 } 351 if (qualifier == null) { 352 qualifier = HConstants.EMPTY_BYTE_ARRAY; 353 } 354 set.add(qualifier); 355 return this; 356 } 357 358 /** 359 * Get versions of columns only within the specified timestamp range, [minStamp, maxStamp). Note, 360 * default maximum versions to return is 1. If your time range spans more than one version and you 361 * want all versions returned, up the number of versions beyond the default. 362 * @param minStamp minimum timestamp value, inclusive 363 * @param maxStamp maximum timestamp value, exclusive 364 * @see #setMaxVersions() 365 * @see #setMaxVersions(int) 366 */ 367 public Scan setTimeRange(long minStamp, long maxStamp) throws IOException { 368 tr = new TimeRange(minStamp, maxStamp); 369 return this; 370 } 371 372 /** 373 * Get versions of columns with the specified timestamp. Note, default maximum versions to return 374 * is 1. If your time range spans more than one version and you want all versions returned, up the 375 * number of versions beyond the defaut. 376 * @param timestamp version timestamp 377 * @see #setMaxVersions() 378 * @see #setMaxVersions(int) 379 * @deprecated As of release 2.0.0, this will be removed in HBase 3.0.0. Use 380 * {@link #setTimestamp(long)} instead 381 */ 382 @Deprecated 383 public Scan setTimeStamp(long timestamp) throws IOException { 384 return this.setTimestamp(timestamp); 385 } 386 387 /** 388 * Get versions of columns with the specified timestamp. Note, default maximum versions to return 389 * is 1. If your time range spans more than one version and you want all versions returned, up the 390 * number of versions beyond the defaut. 391 * @param timestamp version timestamp 392 * @see #setMaxVersions() 393 * @see #setMaxVersions(int) 394 */ 395 public Scan setTimestamp(long timestamp) { 396 try { 397 tr = new TimeRange(timestamp, timestamp + 1); 398 } catch (Exception e) { 399 // This should never happen, unless integer overflow or something extremely wrong... 400 LOG.error("TimeRange failed, likely caused by integer overflow. ", e); 401 throw e; 402 } 403 404 return this; 405 } 406 407 @Override 408 public Scan setColumnFamilyTimeRange(byte[] cf, long minStamp, long maxStamp) { 409 return (Scan) super.setColumnFamilyTimeRange(cf, minStamp, maxStamp); 410 } 411 412 /** 413 * Set the start row of the scan. 414 * <p> 415 * If the specified row does not exist, the Scanner will start from the next closest row after the 416 * specified row. 417 * <p> 418 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 419 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 420 * unexpected or even undefined. 421 * </p> 422 * @param startRow row to start scanner at or after 423 * @throws IllegalArgumentException if startRow does not meet criteria for a row key (when length 424 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 425 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use {@link #withStartRow(byte[])} 426 * instead. This method may change the inclusive of the stop row to keep compatible 427 * with the old behavior. 428 * @see #withStartRow(byte[]) 429 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 430 */ 431 @Deprecated 432 public Scan setStartRow(byte[] startRow) { 433 withStartRow(startRow); 434 if (ClientUtil.areScanStartRowAndStopRowEqual(this.startRow, this.stopRow)) { 435 // for keeping the old behavior that a scan with the same start and stop row is a get scan. 436 this.includeStopRow = true; 437 } 438 return this; 439 } 440 441 /** 442 * Set the start row of the scan. 443 * <p> 444 * If the specified row does not exist, the Scanner will start from the next closest row after the 445 * specified row. 446 * @param startRow row to start scanner at or after 447 * @throws IllegalArgumentException if startRow does not meet criteria for a row key (when length 448 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 449 */ 450 public Scan withStartRow(byte[] startRow) { 451 return withStartRow(startRow, true); 452 } 453 454 /** 455 * Set the start row of the scan. 456 * <p> 457 * If the specified row does not exist, or the {@code inclusive} is {@code false}, the Scanner 458 * will start from the next closest row after the specified row. 459 * <p> 460 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 461 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 462 * unexpected or even undefined. 463 * </p> 464 * @param startRow row to start scanner at or after 465 * @param inclusive whether we should include the start row when scan 466 * @throws IllegalArgumentException if startRow does not meet criteria for a row key (when length 467 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 468 */ 469 public Scan withStartRow(byte[] startRow, boolean inclusive) { 470 if (Bytes.len(startRow) > HConstants.MAX_ROW_LENGTH) { 471 throw new IllegalArgumentException("startRow's length must be less than or equal to " 472 + HConstants.MAX_ROW_LENGTH + " to meet the criteria" + " for a row key."); 473 } 474 this.startRow = startRow; 475 this.includeStartRow = inclusive; 476 return this; 477 } 478 479 /** 480 * Set the stop row of the scan. 481 * <p> 482 * The scan will include rows that are lexicographically less than the provided stopRow. 483 * <p> 484 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 485 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 486 * unexpected or even undefined. 487 * </p> 488 * @param stopRow row to end at (exclusive) 489 * @throws IllegalArgumentException if stopRow does not meet criteria for a row key (when length 490 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 491 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use {@link #withStopRow(byte[])} instead. 492 * This method may change the inclusive of the stop row to keep compatible with the 493 * old behavior. 494 * @see #withStopRow(byte[]) 495 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 496 */ 497 @Deprecated 498 public Scan setStopRow(byte[] stopRow) { 499 withStopRow(stopRow); 500 if (ClientUtil.areScanStartRowAndStopRowEqual(this.startRow, this.stopRow)) { 501 // for keeping the old behavior that a scan with the same start and stop row is a get scan. 502 this.includeStopRow = true; 503 } 504 return this; 505 } 506 507 /** 508 * Set the stop row of the scan. 509 * <p> 510 * The scan will include rows that are lexicographically less than the provided stopRow. 511 * <p> 512 * <b>Note:</b> When doing a filter for a rowKey <u>Prefix</u> use 513 * {@link #setRowPrefixFilter(byte[])}. The 'trailing 0' will not yield the desired result. 514 * </p> 515 * @param stopRow row to end at (exclusive) 516 * @throws IllegalArgumentException if stopRow does not meet criteria for a row key (when length 517 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 518 */ 519 public Scan withStopRow(byte[] stopRow) { 520 return withStopRow(stopRow, false); 521 } 522 523 /** 524 * Set the stop row of the scan. 525 * <p> 526 * The scan will include rows that are lexicographically less than (or equal to if 527 * {@code inclusive} is {@code true}) the provided stopRow. 528 * <p> 529 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 530 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 531 * unexpected or even undefined. 532 * </p> 533 * @param stopRow row to end at 534 * @param inclusive whether we should include the stop row when scan 535 * @throws IllegalArgumentException if stopRow does not meet criteria for a row key (when length 536 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 537 */ 538 public Scan withStopRow(byte[] stopRow, boolean inclusive) { 539 if (Bytes.len(stopRow) > HConstants.MAX_ROW_LENGTH) { 540 throw new IllegalArgumentException("stopRow's length must be less than or equal to " 541 + HConstants.MAX_ROW_LENGTH + " to meet the criteria" + " for a row key."); 542 } 543 this.stopRow = stopRow; 544 this.includeStopRow = inclusive; 545 return this; 546 } 547 548 /** 549 * <p> 550 * Set a filter (using stopRow and startRow) so the result set only contains rows where the rowKey 551 * starts with the specified prefix. 552 * </p> 553 * <p> 554 * This is a utility method that converts the desired rowPrefix into the appropriate values for 555 * the startRow and stopRow to achieve the desired result. 556 * </p> 557 * <p> 558 * This can safely be used in combination with setFilter. 559 * </p> 560 * <p> 561 * <strong>This CANNOT be used in combination with withStartRow and/or withStopRow.</strong> Such 562 * a combination will yield unexpected and even undefined results. 563 * </p> 564 * @param rowPrefix the prefix all rows must start with. (Set <i>null</i> to remove the filter.) 565 * @deprecated since 2.5.0, will be removed in 4.0.0. The name of this method is considered to be 566 * confusing as it does not use a {@link Filter} but uses setting the startRow and 567 * stopRow instead. Use {@link #setStartStopRowForPrefixScan(byte[])} instead. 568 */ 569 public Scan setRowPrefixFilter(byte[] rowPrefix) { 570 return setStartStopRowForPrefixScan(rowPrefix); 571 } 572 573 /** 574 * <p> 575 * Set a filter (using stopRow and startRow) so the result set only contains rows where the rowKey 576 * starts with the specified prefix. 577 * </p> 578 * <p> 579 * This is a utility method that converts the desired rowPrefix into the appropriate values for 580 * the startRow and stopRow to achieve the desired result. 581 * </p> 582 * <p> 583 * This can safely be used in combination with setFilter. 584 * </p> 585 * <p> 586 * <strong>This CANNOT be used in combination with withStartRow and/or withStopRow.</strong> Such 587 * a combination will yield unexpected and even undefined results. 588 * </p> 589 * @param rowPrefix the prefix all rows must start with. (Set <i>null</i> to remove the filter.) 590 */ 591 public Scan setStartStopRowForPrefixScan(byte[] rowPrefix) { 592 if (rowPrefix == null) { 593 setStartRow(HConstants.EMPTY_START_ROW); 594 setStopRow(HConstants.EMPTY_END_ROW); 595 } else { 596 this.setStartRow(rowPrefix); 597 this.setStopRow(ClientUtil.calculateTheClosestNextRowKeyForPrefix(rowPrefix)); 598 } 599 return this; 600 } 601 602 /** 603 * Get all available versions. 604 * @deprecated since 2.0.0 and will be removed in 3.0.0. It is easy to misunderstand with column 605 * family's max versions, so use {@link #readAllVersions()} instead. 606 * @see #readAllVersions() 607 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17125">HBASE-17125</a> 608 */ 609 @Deprecated 610 public Scan setMaxVersions() { 611 return readAllVersions(); 612 } 613 614 /** 615 * Get up to the specified number of versions of each column. 616 * @param maxVersions maximum versions for each column 617 * @deprecated since 2.0.0 and will be removed in 3.0.0. It is easy to misunderstand with column 618 * family's max versions, so use {@link #readVersions(int)} instead. 619 * @see #readVersions(int) 620 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17125">HBASE-17125</a> 621 */ 622 @Deprecated 623 public Scan setMaxVersions(int maxVersions) { 624 return readVersions(maxVersions); 625 } 626 627 /** 628 * Get all available versions. 629 */ 630 public Scan readAllVersions() { 631 this.maxVersions = Integer.MAX_VALUE; 632 return this; 633 } 634 635 /** 636 * Get up to the specified number of versions of each column. 637 * @param versions specified number of versions for each column 638 */ 639 public Scan readVersions(int versions) { 640 this.maxVersions = versions; 641 return this; 642 } 643 644 /** 645 * Set the maximum number of cells to return for each call to next(). Callers should be aware that 646 * this is not equivalent to calling {@link #setAllowPartialResults(boolean)}. If you don't allow 647 * partial results, the number of cells in each Result must equal to your batch setting unless it 648 * is the last Result for current row. So this method is helpful in paging queries. If you just 649 * want to prevent OOM at client, use setAllowPartialResults(true) is better. 650 * @param batch the maximum number of values 651 * @see Result#mayHaveMoreCellsInRow() 652 */ 653 public Scan setBatch(int batch) { 654 if (this.hasFilter() && this.filter.hasFilterRow()) { 655 throw new IncompatibleFilterException( 656 "Cannot set batch on a scan using a filter" + " that returns true for filter.hasFilterRow"); 657 } 658 this.batch = batch; 659 return this; 660 } 661 662 /** 663 * Set the maximum number of values to return per row per Column Family 664 * @param limit the maximum number of values returned / row / CF 665 */ 666 public Scan setMaxResultsPerColumnFamily(int limit) { 667 this.storeLimit = limit; 668 return this; 669 } 670 671 /** 672 * Set offset for the row per Column Family. 673 * @param offset is the number of kvs that will be skipped. 674 */ 675 public Scan setRowOffsetPerColumnFamily(int offset) { 676 this.storeOffset = offset; 677 return this; 678 } 679 680 /** 681 * Set the number of rows for caching that will be passed to scanners. If not set, the 682 * Configuration setting {@link HConstants#HBASE_CLIENT_SCANNER_CACHING} will apply. Higher 683 * caching values will enable faster scanners but will use more memory. 684 * @param caching the number of rows for caching 685 */ 686 public Scan setCaching(int caching) { 687 this.caching = caching; 688 return this; 689 } 690 691 /** Returns the maximum result size in bytes. See {@link #setMaxResultSize(long)} */ 692 public long getMaxResultSize() { 693 return maxResultSize; 694 } 695 696 /** 697 * Set the maximum result size. The default is -1; this means that no specific maximum result size 698 * will be set for this scan, and the global configured value will be used instead. (Defaults to 699 * unlimited). 700 * @param maxResultSize The maximum result size in bytes. 701 */ 702 public Scan setMaxResultSize(long maxResultSize) { 703 this.maxResultSize = maxResultSize; 704 return this; 705 } 706 707 @Override 708 public Scan setFilter(Filter filter) { 709 super.setFilter(filter); 710 return this; 711 } 712 713 /** 714 * Setting the familyMap 715 * @param familyMap map of family to qualifier 716 */ 717 public Scan setFamilyMap(Map<byte[], NavigableSet<byte[]>> familyMap) { 718 this.familyMap = familyMap; 719 return this; 720 } 721 722 /** 723 * Getting the familyMap 724 */ 725 public Map<byte[], NavigableSet<byte[]>> getFamilyMap() { 726 return this.familyMap; 727 } 728 729 /** Returns the number of families in familyMap */ 730 public int numFamilies() { 731 if (hasFamilies()) { 732 return this.familyMap.size(); 733 } 734 return 0; 735 } 736 737 /** Returns true if familyMap is non empty, false otherwise */ 738 public boolean hasFamilies() { 739 return !this.familyMap.isEmpty(); 740 } 741 742 /** Returns the keys of the familyMap */ 743 public byte[][] getFamilies() { 744 if (hasFamilies()) { 745 return this.familyMap.keySet().toArray(new byte[0][0]); 746 } 747 return null; 748 } 749 750 /** Returns the startrow */ 751 public byte[] getStartRow() { 752 return this.startRow; 753 } 754 755 /** Returns if we should include start row when scan */ 756 public boolean includeStartRow() { 757 return includeStartRow; 758 } 759 760 /** Returns the stoprow */ 761 public byte[] getStopRow() { 762 return this.stopRow; 763 } 764 765 /** Returns if we should include stop row when scan */ 766 public boolean includeStopRow() { 767 return includeStopRow; 768 } 769 770 /** Returns the max number of versions to fetch */ 771 public int getMaxVersions() { 772 return this.maxVersions; 773 } 774 775 /** Returns maximum number of values to return for a single call to next() */ 776 public int getBatch() { 777 return this.batch; 778 } 779 780 /** Returns maximum number of values to return per row per CF */ 781 public int getMaxResultsPerColumnFamily() { 782 return this.storeLimit; 783 } 784 785 /** 786 * Method for retrieving the scan's offset per row per column family (#kvs to be skipped) 787 * @return row offset 788 */ 789 public int getRowOffsetPerColumnFamily() { 790 return this.storeOffset; 791 } 792 793 /** Returns caching the number of rows fetched when calling next on a scanner */ 794 public int getCaching() { 795 return this.caching; 796 } 797 798 /** Returns TimeRange */ 799 public TimeRange getTimeRange() { 800 return this.tr; 801 } 802 803 /** Returns RowFilter */ 804 @Override 805 public Filter getFilter() { 806 return filter; 807 } 808 809 /** Returns true is a filter has been specified, false if not */ 810 public boolean hasFilter() { 811 return filter != null; 812 } 813 814 /** 815 * Set whether blocks should be cached for this Scan. 816 * <p> 817 * This is true by default. When true, default settings of the table and family are used (this 818 * will never override caching blocks if the block cache is disabled for that family or entirely). 819 * @param cacheBlocks if false, default settings are overridden and blocks will not be cached 820 */ 821 public Scan setCacheBlocks(boolean cacheBlocks) { 822 this.cacheBlocks = cacheBlocks; 823 return this; 824 } 825 826 /** 827 * Get whether blocks should be cached for this Scan. 828 * @return true if default caching should be used, false if blocks should not be cached 829 */ 830 public boolean getCacheBlocks() { 831 return cacheBlocks; 832 } 833 834 /** 835 * Set whether this scan is a reversed one 836 * <p> 837 * This is false by default which means forward(normal) scan. 838 * @param reversed if true, scan will be backward order 839 */ 840 public Scan setReversed(boolean reversed) { 841 this.reversed = reversed; 842 return this; 843 } 844 845 /** 846 * Get whether this scan is a reversed one. 847 * @return true if backward scan, false if forward(default) scan 848 */ 849 public boolean isReversed() { 850 return reversed; 851 } 852 853 /** 854 * Setting whether the caller wants to see the partial results when server returns 855 * less-than-expected cells. It is helpful while scanning a huge row to prevent OOM at client. By 856 * default this value is false and the complete results will be assembled client side before being 857 * delivered to the caller. 858 * @see Result#mayHaveMoreCellsInRow() 859 * @see #setBatch(int) 860 */ 861 public Scan setAllowPartialResults(final boolean allowPartialResults) { 862 this.allowPartialResults = allowPartialResults; 863 return this; 864 } 865 866 /** 867 * Returns true when the constructor of this scan understands that the results they will see may 868 * only represent a partial portion of a row. The entire row would be retrieved by subsequent 869 * calls to {@link ResultScanner#next()} 870 */ 871 public boolean getAllowPartialResults() { 872 return allowPartialResults; 873 } 874 875 @Override 876 public Scan setLoadColumnFamiliesOnDemand(boolean value) { 877 return (Scan) super.setLoadColumnFamiliesOnDemand(value); 878 } 879 880 /** 881 * Compile the table and column family (i.e. schema) information into a String. Useful for parsing 882 * and aggregation by debugging, logging, and administration tools. 883 */ 884 @Override 885 public Map<String, Object> getFingerprint() { 886 Map<String, Object> map = new HashMap<>(); 887 List<String> families = new ArrayList<>(); 888 if (this.familyMap.isEmpty()) { 889 map.put("families", "ALL"); 890 return map; 891 } else { 892 map.put("families", families); 893 } 894 for (Map.Entry<byte[], NavigableSet<byte[]>> entry : this.familyMap.entrySet()) { 895 families.add(Bytes.toStringBinary(entry.getKey())); 896 } 897 return map; 898 } 899 900 /** 901 * Compile the details beyond the scope of getFingerprint (row, columns, timestamps, etc.) into a 902 * Map along with the fingerprinted information. Useful for debugging, logging, and administration 903 * tools. 904 * @param maxCols a limit on the number of columns output prior to truncation 905 */ 906 @Override 907 public Map<String, Object> toMap(int maxCols) { 908 // start with the fingerprint map and build on top of it 909 Map<String, Object> map = getFingerprint(); 910 // map from families to column list replaces fingerprint's list of families 911 Map<String, List<String>> familyColumns = new HashMap<>(); 912 map.put("families", familyColumns); 913 // add scalar information first 914 map.put("startRow", Bytes.toStringBinary(this.startRow)); 915 map.put("stopRow", Bytes.toStringBinary(this.stopRow)); 916 map.put("maxVersions", this.maxVersions); 917 map.put("batch", this.batch); 918 map.put("caching", this.caching); 919 map.put("maxResultSize", this.maxResultSize); 920 map.put("cacheBlocks", this.cacheBlocks); 921 map.put("loadColumnFamiliesOnDemand", this.loadColumnFamiliesOnDemand); 922 List<Long> timeRange = new ArrayList<>(2); 923 timeRange.add(this.tr.getMin()); 924 timeRange.add(this.tr.getMax()); 925 map.put("timeRange", timeRange); 926 int colCount = 0; 927 // iterate through affected families and list out up to maxCols columns 928 for (Map.Entry<byte[], NavigableSet<byte[]>> entry : this.familyMap.entrySet()) { 929 List<String> columns = new ArrayList<>(); 930 familyColumns.put(Bytes.toStringBinary(entry.getKey()), columns); 931 if (entry.getValue() == null) { 932 colCount++; 933 --maxCols; 934 columns.add("ALL"); 935 } else { 936 colCount += entry.getValue().size(); 937 if (maxCols <= 0) { 938 continue; 939 } 940 for (byte[] column : entry.getValue()) { 941 if (--maxCols <= 0) { 942 continue; 943 } 944 columns.add(Bytes.toStringBinary(column)); 945 } 946 } 947 } 948 map.put("totalColumns", colCount); 949 if (this.filter != null) { 950 map.put("filter", this.filter.toString()); 951 } 952 // add the id if set 953 if (getId() != null) { 954 map.put("id", getId()); 955 } 956 map.put("includeStartRow", includeStartRow); 957 map.put("includeStopRow", includeStopRow); 958 map.put("allowPartialResults", allowPartialResults); 959 map.put("storeLimit", storeLimit); 960 map.put("storeOffset", storeOffset); 961 map.put("reversed", reversed); 962 if (null != asyncPrefetch) { 963 map.put("asyncPrefetch", asyncPrefetch); 964 } 965 map.put("mvccReadPoint", mvccReadPoint); 966 map.put("limit", limit); 967 map.put("readType", readType); 968 map.put("needCursorResult", needCursorResult); 969 map.put("targetReplicaId", targetReplicaId); 970 map.put("consistency", consistency); 971 if (!colFamTimeRangeMap.isEmpty()) { 972 Map<String, List<Long>> colFamTimeRangeMapStr = colFamTimeRangeMap.entrySet().stream() 973 .collect(Collectors.toMap((e) -> Bytes.toStringBinary(e.getKey()), e -> { 974 TimeRange value = e.getValue(); 975 List<Long> rangeList = new ArrayList<>(); 976 rangeList.add(value.getMin()); 977 rangeList.add(value.getMax()); 978 return rangeList; 979 })); 980 981 map.put("colFamTimeRangeMap", colFamTimeRangeMapStr); 982 } 983 map.put("priority", getPriority()); 984 return map; 985 } 986 987 /** 988 * Enable/disable "raw" mode for this scan. If "raw" is enabled the scan will return all delete 989 * marker and deleted rows that have not been collected, yet. This is mostly useful for Scan on 990 * column families that have KEEP_DELETED_ROWS enabled. It is an error to specify any column when 991 * "raw" is set. 992 * @param raw True/False to enable/disable "raw" mode. 993 */ 994 public Scan setRaw(boolean raw) { 995 setAttribute(RAW_ATTR, Bytes.toBytes(raw)); 996 return this; 997 } 998 999 /** Returns True if this Scan is in "raw" mode. */ 1000 public boolean isRaw() { 1001 byte[] attr = getAttribute(RAW_ATTR); 1002 return attr == null ? false : Bytes.toBoolean(attr); 1003 } 1004 1005 /** 1006 * Set whether this scan is a small scan 1007 * <p> 1008 * Small scan should use pread and big scan can use seek + read seek + read is fast but can cause 1009 * two problem (1) resource contention (2) cause too much network io [89-fb] Using pread for 1010 * non-compaction read request https://issues.apache.org/jira/browse/HBASE-7266 On the other hand, 1011 * if setting it true, we would do openScanner,next,closeScanner in one RPC call. It means the 1012 * better performance for small scan. [HBASE-9488]. Generally, if the scan range is within one 1013 * data block(64KB), it could be considered as a small scan. 1014 * @param small set if that should use read type of PREAD 1015 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use {@link #setLimit(int)} and 1016 * {@link #setReadType(ReadType)} instead. And for the one rpc optimization, now we 1017 * will also fetch data when openScanner, and if the number of rows reaches the limit 1018 * then we will close the scanner automatically which means we will fall back to one 1019 * rpc. 1020 * @see #setLimit(int) 1021 * @see #setReadType(ReadType) 1022 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17045">HBASE-17045</a> 1023 */ 1024 @Deprecated 1025 public Scan setSmall(boolean small) { 1026 this.small = small; 1027 if (small) { 1028 this.readType = ReadType.PREAD; 1029 } 1030 return this; 1031 } 1032 1033 /** 1034 * Get whether this scan is a small scan 1035 * @return true if small scan 1036 * @deprecated since 2.0.0 and will be removed in 3.0.0. See the comment of 1037 * {@link #setSmall(boolean)} 1038 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17045">HBASE-17045</a> 1039 */ 1040 @Deprecated 1041 public boolean isSmall() { 1042 return small; 1043 } 1044 1045 @Override 1046 public Scan setAttribute(String name, byte[] value) { 1047 return (Scan) super.setAttribute(name, value); 1048 } 1049 1050 @Override 1051 public Scan setId(String id) { 1052 return (Scan) super.setId(id); 1053 } 1054 1055 @Override 1056 public Scan setAuthorizations(Authorizations authorizations) { 1057 return (Scan) super.setAuthorizations(authorizations); 1058 } 1059 1060 @Override 1061 public Scan setACL(Map<String, Permission> perms) { 1062 return (Scan) super.setACL(perms); 1063 } 1064 1065 @Override 1066 public Scan setACL(String user, Permission perms) { 1067 return (Scan) super.setACL(user, perms); 1068 } 1069 1070 @Override 1071 public Scan setConsistency(Consistency consistency) { 1072 return (Scan) super.setConsistency(consistency); 1073 } 1074 1075 @Override 1076 public Scan setReplicaId(int Id) { 1077 return (Scan) super.setReplicaId(Id); 1078 } 1079 1080 @Override 1081 public Scan setIsolationLevel(IsolationLevel level) { 1082 return (Scan) super.setIsolationLevel(level); 1083 } 1084 1085 @Override 1086 public Scan setPriority(int priority) { 1087 return (Scan) super.setPriority(priority); 1088 } 1089 1090 /** 1091 * Enable collection of {@link ScanMetrics}. For advanced users. 1092 * @param enabled Set to true to enable accumulating scan metrics 1093 */ 1094 public Scan setScanMetricsEnabled(final boolean enabled) { 1095 setAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE, Bytes.toBytes(Boolean.valueOf(enabled))); 1096 return this; 1097 } 1098 1099 /** Returns True if collection of scan metrics is enabled. For advanced users. */ 1100 public boolean isScanMetricsEnabled() { 1101 byte[] attr = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE); 1102 return attr == null ? false : Bytes.toBoolean(attr); 1103 } 1104 1105 /** 1106 * @return Metrics on this Scan, if metrics were enabled. 1107 * @see #setScanMetricsEnabled(boolean) 1108 * @deprecated Use {@link ResultScanner#getScanMetrics()} instead. And notice that, please do not 1109 * use this method and {@link ResultScanner#getScanMetrics()} together, the metrics 1110 * will be messed up. 1111 */ 1112 @Deprecated 1113 public ScanMetrics getScanMetrics() { 1114 byte[] bytes = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_DATA); 1115 if (bytes == null) return null; 1116 return ProtobufUtil.toScanMetrics(bytes); 1117 } 1118 1119 public Boolean isAsyncPrefetch() { 1120 return asyncPrefetch; 1121 } 1122 1123 public Scan setAsyncPrefetch(boolean asyncPrefetch) { 1124 this.asyncPrefetch = asyncPrefetch; 1125 return this; 1126 } 1127 1128 /** Returns the limit of rows for this scan */ 1129 public int getLimit() { 1130 return limit; 1131 } 1132 1133 /** 1134 * Set the limit of rows for this scan. We will terminate the scan if the number of returned rows 1135 * reaches this value. 1136 * <p> 1137 * This condition will be tested at last, after all other conditions such as stopRow, filter, etc. 1138 * @param limit the limit of rows for this scan 1139 */ 1140 public Scan setLimit(int limit) { 1141 this.limit = limit; 1142 return this; 1143 } 1144 1145 /** 1146 * Call this when you only want to get one row. It will set {@code limit} to {@code 1}, and also 1147 * set {@code readType} to {@link ReadType#PREAD}. 1148 */ 1149 public Scan setOneRowLimit() { 1150 return setLimit(1).setReadType(ReadType.PREAD); 1151 } 1152 1153 @InterfaceAudience.Public 1154 public enum ReadType { 1155 DEFAULT, 1156 STREAM, 1157 PREAD 1158 } 1159 1160 /** Returns the read type for this scan */ 1161 public ReadType getReadType() { 1162 return readType; 1163 } 1164 1165 /** 1166 * Set the read type for this scan. 1167 * <p> 1168 * Notice that we may choose to use pread even if you specific {@link ReadType#STREAM} here. For 1169 * example, we will always use pread if this is a get scan. 1170 */ 1171 public Scan setReadType(ReadType readType) { 1172 this.readType = readType; 1173 return this; 1174 } 1175 1176 /** 1177 * Get the mvcc read point used to open a scanner. 1178 */ 1179 long getMvccReadPoint() { 1180 return mvccReadPoint; 1181 } 1182 1183 /** 1184 * Set the mvcc read point used to open a scanner. 1185 */ 1186 Scan setMvccReadPoint(long mvccReadPoint) { 1187 this.mvccReadPoint = mvccReadPoint; 1188 return this; 1189 } 1190 1191 /** 1192 * Set the mvcc read point to -1 which means do not use it. 1193 */ 1194 Scan resetMvccReadPoint() { 1195 return setMvccReadPoint(-1L); 1196 } 1197 1198 /** 1199 * When the server is slow or we scan a table with many deleted data or we use a sparse filter, 1200 * the server will response heartbeat to prevent timeout. However the scanner will return a Result 1201 * only when client can do it. So if there are many heartbeats, the blocking time on 1202 * ResultScanner#next() may be very long, which is not friendly to online services. Set this to 1203 * true then you can get a special Result whose #isCursor() returns true and is not contains any 1204 * real data. It only tells you where the server has scanned. You can call next to continue 1205 * scanning or open a new scanner with this row key as start row whenever you want. Users can get 1206 * a cursor when and only when there is a response from the server but we can not return a Result 1207 * to users, for example, this response is a heartbeat or there are partial cells but users do not 1208 * allow partial result. Now the cursor is in row level which means the special Result will only 1209 * contains a row key. {@link Result#isCursor()} {@link Result#getCursor()} {@link Cursor} 1210 */ 1211 public Scan setNeedCursorResult(boolean needCursorResult) { 1212 this.needCursorResult = needCursorResult; 1213 return this; 1214 } 1215 1216 public boolean isNeedCursorResult() { 1217 return needCursorResult; 1218 } 1219 1220 /** 1221 * Create a new Scan with a cursor. It only set the position information like start row key. The 1222 * others (like cfs, stop row, limit) should still be filled in by the user. 1223 * {@link Result#isCursor()} {@link Result#getCursor()} {@link Cursor} 1224 */ 1225 public static Scan createScanFromCursor(Cursor cursor) { 1226 return new Scan().withStartRow(cursor.getRow()); 1227 } 1228}