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 static private final String SCAN_ATTRIBUTES_METRICS_BY_REGION_ENABLE = 134 "scan.attributes.metrics.byregion.enable"; 135 136 /** 137 * -1 means no caching specified and the value of {@link HConstants#HBASE_CLIENT_SCANNER_CACHING} 138 * (default to {@link HConstants#DEFAULT_HBASE_CLIENT_SCANNER_CACHING}) will be used 139 */ 140 private int caching = -1; 141 private long maxResultSize = -1; 142 private boolean cacheBlocks = true; 143 private boolean reversed = false; 144 private TimeRange tr = TimeRange.allTime(); 145 private Map<byte[], NavigableSet<byte[]>> familyMap = 146 new TreeMap<byte[], NavigableSet<byte[]>>(Bytes.BYTES_COMPARATOR); 147 private Boolean asyncPrefetch = null; 148 149 /** 150 * Parameter name for client scanner sync/async prefetch toggle. When using async scanner, 151 * prefetching data from the server is done at the background. The parameter currently won't have 152 * any effect in the case that the user has set Scan#setSmall or Scan#setReversed 153 */ 154 public static final String HBASE_CLIENT_SCANNER_ASYNC_PREFETCH = 155 "hbase.client.scanner.async.prefetch"; 156 157 /** 158 * Default value of {@link #HBASE_CLIENT_SCANNER_ASYNC_PREFETCH}. 159 */ 160 public static final boolean DEFAULT_HBASE_CLIENT_SCANNER_ASYNC_PREFETCH = false; 161 162 /** 163 * Set it true for small scan to get better performance Small scan should use pread and big scan 164 * can use seek + read seek + read is fast but can cause two problem (1) resource contention (2) 165 * cause too much network io [89-fb] Using pread for non-compaction read request 166 * https://issues.apache.org/jira/browse/HBASE-7266 On the other hand, if setting it true, we 167 * would do openScanner,next,closeScanner in one RPC call. It means the better performance for 168 * small scan. [HBASE-9488]. Generally, if the scan range is within one data block(64KB), it could 169 * be considered as a small scan. 170 */ 171 private boolean small = false; 172 173 /** 174 * The mvcc read point to use when open a scanner. Remember to clear it after switching regions as 175 * the mvcc is only valid within region scope. 176 */ 177 private long mvccReadPoint = -1L; 178 179 /** 180 * The number of rows we want for this scan. We will terminate the scan if the number of return 181 * rows reaches this value. 182 */ 183 private int limit = -1; 184 185 /** 186 * Control whether to use pread at server side. 187 */ 188 private ReadType readType = ReadType.DEFAULT; 189 190 private boolean needCursorResult = false; 191 192 /** 193 * Create a Scan operation across all rows. 194 */ 195 public Scan() { 196 } 197 198 /** 199 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use 200 * {@code new Scan().withStartRow(startRow).setFilter(filter)} instead. 201 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 202 */ 203 @Deprecated 204 public Scan(byte[] startRow, Filter filter) { 205 this(startRow); 206 this.filter = filter; 207 } 208 209 /** 210 * Create a Scan operation starting at the specified row. 211 * <p> 212 * If the specified row does not exist, the Scanner will start from the next closest row after the 213 * specified row. 214 * @param startRow row to start scanner at or after 215 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use 216 * {@code new Scan().withStartRow(startRow)} instead. 217 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 218 */ 219 @Deprecated 220 public Scan(byte[] startRow) { 221 setStartRow(startRow); 222 } 223 224 /** 225 * Create a Scan operation for the range of rows specified. 226 * @param startRow row to start scanner at or after (inclusive) 227 * @param stopRow row to stop scanner before (exclusive) 228 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use 229 * {@code new Scan().withStartRow(startRow).withStopRow(stopRow)} instead. 230 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 231 */ 232 @Deprecated 233 public Scan(byte[] startRow, byte[] stopRow) { 234 setStartRow(startRow); 235 setStopRow(stopRow); 236 } 237 238 /** 239 * Creates a new instance of this class while copying all values. 240 * @param scan The scan instance to copy from. 241 * @throws IOException When copying the values fails. 242 */ 243 public Scan(Scan scan) throws IOException { 244 startRow = scan.getStartRow(); 245 includeStartRow = scan.includeStartRow(); 246 stopRow = scan.getStopRow(); 247 includeStopRow = scan.includeStopRow(); 248 maxVersions = scan.getMaxVersions(); 249 batch = scan.getBatch(); 250 storeLimit = scan.getMaxResultsPerColumnFamily(); 251 storeOffset = scan.getRowOffsetPerColumnFamily(); 252 caching = scan.getCaching(); 253 maxResultSize = scan.getMaxResultSize(); 254 cacheBlocks = scan.getCacheBlocks(); 255 filter = scan.getFilter(); // clone? 256 loadColumnFamiliesOnDemand = scan.getLoadColumnFamiliesOnDemandValue(); 257 consistency = scan.getConsistency(); 258 this.setIsolationLevel(scan.getIsolationLevel()); 259 reversed = scan.isReversed(); 260 asyncPrefetch = scan.isAsyncPrefetch(); 261 small = scan.isSmall(); 262 allowPartialResults = scan.getAllowPartialResults(); 263 tr = scan.getTimeRange(); // TimeRange is immutable 264 Map<byte[], NavigableSet<byte[]>> fams = scan.getFamilyMap(); 265 for (Map.Entry<byte[], NavigableSet<byte[]>> entry : fams.entrySet()) { 266 byte[] fam = entry.getKey(); 267 NavigableSet<byte[]> cols = entry.getValue(); 268 if (cols != null && cols.size() > 0) { 269 for (byte[] col : cols) { 270 addColumn(fam, col); 271 } 272 } else { 273 addFamily(fam); 274 } 275 } 276 for (Map.Entry<String, byte[]> attr : scan.getAttributesMap().entrySet()) { 277 setAttribute(attr.getKey(), attr.getValue()); 278 } 279 for (Map.Entry<byte[], TimeRange> entry : scan.getColumnFamilyTimeRange().entrySet()) { 280 TimeRange tr = entry.getValue(); 281 setColumnFamilyTimeRange(entry.getKey(), tr.getMin(), tr.getMax()); 282 } 283 this.mvccReadPoint = scan.getMvccReadPoint(); 284 this.limit = scan.getLimit(); 285 this.needCursorResult = scan.isNeedCursorResult(); 286 setPriority(scan.getPriority()); 287 readType = scan.getReadType(); 288 super.setReplicaId(scan.getReplicaId()); 289 super.setQueryMetricsEnabled(scan.isQueryMetricsEnabled()); 290 } 291 292 /** 293 * Builds a scan object with the same specs as get. 294 * @param get get to model scan after 295 */ 296 public Scan(Get get) { 297 this.startRow = get.getRow(); 298 this.includeStartRow = true; 299 this.stopRow = get.getRow(); 300 this.includeStopRow = true; 301 this.filter = get.getFilter(); 302 this.cacheBlocks = get.getCacheBlocks(); 303 this.maxVersions = get.getMaxVersions(); 304 this.storeLimit = get.getMaxResultsPerColumnFamily(); 305 this.storeOffset = get.getRowOffsetPerColumnFamily(); 306 this.tr = get.getTimeRange(); 307 this.familyMap = get.getFamilyMap(); 308 this.asyncPrefetch = false; 309 this.consistency = get.getConsistency(); 310 this.setIsolationLevel(get.getIsolationLevel()); 311 this.loadColumnFamiliesOnDemand = get.getLoadColumnFamiliesOnDemandValue(); 312 for (Map.Entry<String, byte[]> attr : get.getAttributesMap().entrySet()) { 313 setAttribute(attr.getKey(), attr.getValue()); 314 } 315 for (Map.Entry<byte[], TimeRange> entry : get.getColumnFamilyTimeRange().entrySet()) { 316 TimeRange tr = entry.getValue(); 317 setColumnFamilyTimeRange(entry.getKey(), tr.getMin(), tr.getMax()); 318 } 319 this.mvccReadPoint = -1L; 320 setPriority(get.getPriority()); 321 super.setReplicaId(get.getReplicaId()); 322 super.setQueryMetricsEnabled(get.isQueryMetricsEnabled()); 323 } 324 325 public boolean isGetScan() { 326 return includeStartRow && includeStopRow 327 && ClientUtil.areScanStartRowAndStopRowEqual(this.startRow, this.stopRow); 328 } 329 330 /** 331 * Get all columns from the specified family. 332 * <p> 333 * Overrides previous calls to addColumn for this family. 334 * @param family family name 335 */ 336 public Scan addFamily(byte[] family) { 337 familyMap.remove(family); 338 familyMap.put(family, null); 339 return this; 340 } 341 342 /** 343 * Get the column from the specified family with the specified qualifier. 344 * <p> 345 * Overrides previous calls to addFamily for this family. 346 * @param family family name 347 * @param qualifier column qualifier 348 */ 349 public Scan addColumn(byte[] family, byte[] qualifier) { 350 NavigableSet<byte[]> set = familyMap.get(family); 351 if (set == null) { 352 set = new TreeSet<>(Bytes.BYTES_COMPARATOR); 353 familyMap.put(family, set); 354 } 355 if (qualifier == null) { 356 qualifier = HConstants.EMPTY_BYTE_ARRAY; 357 } 358 set.add(qualifier); 359 return this; 360 } 361 362 /** 363 * Get versions of columns only within the specified timestamp range, [minStamp, maxStamp). Note, 364 * default maximum versions to return is 1. If your time range spans more than one version and you 365 * want all versions returned, up the number of versions beyond the default. 366 * @param minStamp minimum timestamp value, inclusive 367 * @param maxStamp maximum timestamp value, exclusive 368 * @see #setMaxVersions() 369 * @see #setMaxVersions(int) 370 */ 371 public Scan setTimeRange(long minStamp, long maxStamp) throws IOException { 372 tr = new TimeRange(minStamp, maxStamp); 373 return this; 374 } 375 376 /** 377 * Get versions of columns with the specified timestamp. Note, default maximum versions to return 378 * is 1. If your time range spans more than one version and you want all versions returned, up the 379 * number of versions beyond the defaut. 380 * @param timestamp version timestamp 381 * @see #setMaxVersions() 382 * @see #setMaxVersions(int) 383 * @deprecated As of release 2.0.0, this will be removed in HBase 3.0.0. Use 384 * {@link #setTimestamp(long)} instead 385 */ 386 @Deprecated 387 public Scan setTimeStamp(long timestamp) throws IOException { 388 return this.setTimestamp(timestamp); 389 } 390 391 /** 392 * Get versions of columns with the specified timestamp. Note, default maximum versions to return 393 * is 1. If your time range spans more than one version and you want all versions returned, up the 394 * number of versions beyond the defaut. 395 * @param timestamp version timestamp 396 * @see #setMaxVersions() 397 * @see #setMaxVersions(int) 398 */ 399 public Scan setTimestamp(long timestamp) { 400 try { 401 tr = new TimeRange(timestamp, timestamp + 1); 402 } catch (Exception e) { 403 // This should never happen, unless integer overflow or something extremely wrong... 404 LOG.error("TimeRange failed, likely caused by integer overflow. ", e); 405 throw e; 406 } 407 408 return this; 409 } 410 411 @Override 412 public Scan setColumnFamilyTimeRange(byte[] cf, long minStamp, long maxStamp) { 413 super.setColumnFamilyTimeRange(cf, minStamp, maxStamp); 414 return this; 415 } 416 417 /** 418 * Set the start row of the scan. 419 * <p> 420 * If the specified row does not exist, the Scanner will start from the next closest row after the 421 * specified row. 422 * <p> 423 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 424 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 425 * unexpected or even undefined. 426 * </p> 427 * @param startRow row to start scanner at or after 428 * @throws IllegalArgumentException if startRow does not meet criteria for a row key (when length 429 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 430 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use {@link #withStartRow(byte[])} 431 * instead. This method may change the inclusive of the stop row to keep compatible 432 * with the old behavior. 433 * @see #withStartRow(byte[]) 434 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 435 */ 436 @Deprecated 437 public Scan setStartRow(byte[] startRow) { 438 withStartRow(startRow); 439 if (ClientUtil.areScanStartRowAndStopRowEqual(this.startRow, this.stopRow)) { 440 // for keeping the old behavior that a scan with the same start and stop row is a get scan. 441 this.includeStopRow = true; 442 } 443 return this; 444 } 445 446 /** 447 * Set the start row of the scan. 448 * <p> 449 * If the specified row does not exist, the Scanner will start from the next closest row after the 450 * specified row. 451 * @param startRow row to start scanner at or after 452 * @throws IllegalArgumentException if startRow does not meet criteria for a row key (when length 453 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 454 */ 455 public Scan withStartRow(byte[] startRow) { 456 return withStartRow(startRow, true); 457 } 458 459 /** 460 * Set the start row of the scan. 461 * <p> 462 * If the specified row does not exist, or the {@code inclusive} is {@code false}, the Scanner 463 * will start from the next closest row after the specified row. 464 * <p> 465 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 466 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 467 * unexpected or even undefined. 468 * </p> 469 * @param startRow row to start scanner at or after 470 * @param inclusive whether we should include the start row when scan 471 * @throws IllegalArgumentException if startRow does not meet criteria for a row key (when length 472 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 473 */ 474 public Scan withStartRow(byte[] startRow, boolean inclusive) { 475 if (Bytes.len(startRow) > HConstants.MAX_ROW_LENGTH) { 476 throw new IllegalArgumentException("startRow's length must be less than or equal to " 477 + HConstants.MAX_ROW_LENGTH + " to meet the criteria" + " for a row key."); 478 } 479 this.startRow = startRow; 480 this.includeStartRow = inclusive; 481 return this; 482 } 483 484 /** 485 * Set the stop row of the scan. 486 * <p> 487 * The scan will include rows that are lexicographically less than the provided stopRow. 488 * <p> 489 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 490 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 491 * unexpected or even undefined. 492 * </p> 493 * @param stopRow row to end at (exclusive) 494 * @throws IllegalArgumentException if stopRow does not meet criteria for a row key (when length 495 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 496 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use {@link #withStopRow(byte[])} instead. 497 * This method may change the inclusive of the stop row to keep compatible with the 498 * old behavior. 499 * @see #withStopRow(byte[]) 500 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17320">HBASE-17320</a> 501 */ 502 @Deprecated 503 public Scan setStopRow(byte[] stopRow) { 504 withStopRow(stopRow); 505 if (ClientUtil.areScanStartRowAndStopRowEqual(this.startRow, this.stopRow)) { 506 // for keeping the old behavior that a scan with the same start and stop row is a get scan. 507 this.includeStopRow = true; 508 } 509 return this; 510 } 511 512 /** 513 * Set the stop row of the scan. 514 * <p> 515 * The scan will include rows that are lexicographically less than the provided stopRow. 516 * <p> 517 * <b>Note:</b> When doing a filter for a rowKey <u>Prefix</u> use 518 * {@link #setRowPrefixFilter(byte[])}. The 'trailing 0' will not yield the desired result. 519 * </p> 520 * @param stopRow row to end at (exclusive) 521 * @throws IllegalArgumentException if stopRow does not meet criteria for a row key (when length 522 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 523 */ 524 public Scan withStopRow(byte[] stopRow) { 525 return withStopRow(stopRow, false); 526 } 527 528 /** 529 * Set the stop row of the scan. 530 * <p> 531 * The scan will include rows that are lexicographically less than (or equal to if 532 * {@code inclusive} is {@code true}) the provided stopRow. 533 * <p> 534 * <b>Note:</b> <strong>Do NOT use this in combination with {@link #setRowPrefixFilter(byte[])} or 535 * {@link #setStartStopRowForPrefixScan(byte[])}.</strong> Doing so will make the scan result 536 * unexpected or even undefined. 537 * </p> 538 * @param stopRow row to end at 539 * @param inclusive whether we should include the stop row when scan 540 * @throws IllegalArgumentException if stopRow does not meet criteria for a row key (when length 541 * exceeds {@link HConstants#MAX_ROW_LENGTH}) 542 */ 543 public Scan withStopRow(byte[] stopRow, boolean inclusive) { 544 if (Bytes.len(stopRow) > HConstants.MAX_ROW_LENGTH) { 545 throw new IllegalArgumentException("stopRow's length must be less than or equal to " 546 + HConstants.MAX_ROW_LENGTH + " to meet the criteria" + " for a row key."); 547 } 548 this.stopRow = stopRow; 549 this.includeStopRow = inclusive; 550 return this; 551 } 552 553 /** 554 * <p> 555 * Set a filter (using stopRow and startRow) so the result set only contains rows where the rowKey 556 * starts with the specified prefix. 557 * </p> 558 * <p> 559 * This is a utility method that converts the desired rowPrefix into the appropriate values for 560 * the startRow and stopRow to achieve the desired result. 561 * </p> 562 * <p> 563 * This can safely be used in combination with setFilter. 564 * </p> 565 * <p> 566 * <strong>This CANNOT be used in combination with withStartRow and/or withStopRow.</strong> Such 567 * a combination will yield unexpected and even undefined results. 568 * </p> 569 * @param rowPrefix the prefix all rows must start with. (Set <i>null</i> to remove the filter.) 570 * @deprecated since 2.5.0, will be removed in 4.0.0. The name of this method is considered to be 571 * confusing as it does not use a {@link Filter} but uses setting the startRow and 572 * stopRow instead. Use {@link #setStartStopRowForPrefixScan(byte[])} instead. 573 */ 574 public Scan setRowPrefixFilter(byte[] rowPrefix) { 575 return setStartStopRowForPrefixScan(rowPrefix); 576 } 577 578 /** 579 * <p> 580 * Set a filter (using stopRow and startRow) so the result set only contains rows where the rowKey 581 * starts with the specified prefix. 582 * </p> 583 * <p> 584 * This is a utility method that converts the desired rowPrefix into the appropriate values for 585 * the startRow and stopRow to achieve the desired result. 586 * </p> 587 * <p> 588 * This can safely be used in combination with setFilter. 589 * </p> 590 * <p> 591 * <strong>This CANNOT be used in combination with withStartRow and/or withStopRow.</strong> Such 592 * a combination will yield unexpected and even undefined results. 593 * </p> 594 * @param rowPrefix the prefix all rows must start with. (Set <i>null</i> to remove the filter.) 595 */ 596 public Scan setStartStopRowForPrefixScan(byte[] rowPrefix) { 597 if (rowPrefix == null) { 598 setStartRow(HConstants.EMPTY_START_ROW); 599 setStopRow(HConstants.EMPTY_END_ROW); 600 } else { 601 this.setStartRow(rowPrefix); 602 this.setStopRow(ClientUtil.calculateTheClosestNextRowKeyForPrefix(rowPrefix)); 603 } 604 return this; 605 } 606 607 /** 608 * Get all available versions. 609 * @deprecated since 2.0.0 and will be removed in 3.0.0. It is easy to misunderstand with column 610 * family's max versions, so use {@link #readAllVersions()} instead. 611 * @see #readAllVersions() 612 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17125">HBASE-17125</a> 613 */ 614 @Deprecated 615 public Scan setMaxVersions() { 616 return readAllVersions(); 617 } 618 619 /** 620 * Get up to the specified number of versions of each column. 621 * @param maxVersions maximum versions for each column 622 * @deprecated since 2.0.0 and will be removed in 3.0.0. It is easy to misunderstand with column 623 * family's max versions, so use {@link #readVersions(int)} instead. 624 * @see #readVersions(int) 625 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17125">HBASE-17125</a> 626 */ 627 @Deprecated 628 public Scan setMaxVersions(int maxVersions) { 629 return readVersions(maxVersions); 630 } 631 632 /** 633 * Get all available versions. 634 */ 635 public Scan readAllVersions() { 636 this.maxVersions = Integer.MAX_VALUE; 637 return this; 638 } 639 640 /** 641 * Get up to the specified number of versions of each column. 642 * @param versions specified number of versions for each column 643 */ 644 public Scan readVersions(int versions) { 645 this.maxVersions = versions; 646 return this; 647 } 648 649 /** 650 * Set the maximum number of cells to return for each call to next(). Callers should be aware that 651 * this is not equivalent to calling {@link #setAllowPartialResults(boolean)}. If you don't allow 652 * partial results, the number of cells in each Result must equal to your batch setting unless it 653 * is the last Result for current row. So this method is helpful in paging queries. If you just 654 * want to prevent OOM at client, use setAllowPartialResults(true) is better. 655 * @param batch the maximum number of values 656 * @see Result#mayHaveMoreCellsInRow() 657 */ 658 public Scan setBatch(int batch) { 659 if (this.hasFilter() && this.filter.hasFilterRow()) { 660 throw new IncompatibleFilterException( 661 "Cannot set batch on a scan using a filter" + " that returns true for filter.hasFilterRow"); 662 } 663 this.batch = batch; 664 return this; 665 } 666 667 /** 668 * Set the maximum number of values to return per row per Column Family 669 * @param limit the maximum number of values returned / row / CF 670 */ 671 public Scan setMaxResultsPerColumnFamily(int limit) { 672 this.storeLimit = limit; 673 return this; 674 } 675 676 /** 677 * Set offset for the row per Column Family. 678 * @param offset is the number of kvs that will be skipped. 679 */ 680 public Scan setRowOffsetPerColumnFamily(int offset) { 681 this.storeOffset = offset; 682 return this; 683 } 684 685 /** 686 * Set the number of rows for caching that will be passed to scanners. If not set, the 687 * Configuration setting {@link HConstants#HBASE_CLIENT_SCANNER_CACHING} will apply. Higher 688 * caching values will enable faster scanners but will use more memory. 689 * @param caching the number of rows for caching 690 */ 691 public Scan setCaching(int caching) { 692 this.caching = caching; 693 return this; 694 } 695 696 /** Returns the maximum result size in bytes. See {@link #setMaxResultSize(long)} */ 697 public long getMaxResultSize() { 698 return maxResultSize; 699 } 700 701 /** 702 * Set the maximum result size. The default is -1; this means that no specific maximum result size 703 * will be set for this scan, and the global configured value will be used instead. (Defaults to 704 * unlimited). 705 * @param maxResultSize The maximum result size in bytes. 706 */ 707 public Scan setMaxResultSize(long maxResultSize) { 708 this.maxResultSize = maxResultSize; 709 return this; 710 } 711 712 @Override 713 public Scan setFilter(Filter filter) { 714 super.setFilter(filter); 715 return this; 716 } 717 718 /** 719 * Setting the familyMap 720 * @param familyMap map of family to qualifier 721 */ 722 public Scan setFamilyMap(Map<byte[], NavigableSet<byte[]>> familyMap) { 723 this.familyMap = familyMap; 724 return this; 725 } 726 727 /** 728 * Getting the familyMap 729 */ 730 public Map<byte[], NavigableSet<byte[]>> getFamilyMap() { 731 return this.familyMap; 732 } 733 734 /** Returns the number of families in familyMap */ 735 public int numFamilies() { 736 if (hasFamilies()) { 737 return this.familyMap.size(); 738 } 739 return 0; 740 } 741 742 /** Returns true if familyMap is non empty, false otherwise */ 743 public boolean hasFamilies() { 744 return !this.familyMap.isEmpty(); 745 } 746 747 /** Returns the keys of the familyMap */ 748 public byte[][] getFamilies() { 749 if (hasFamilies()) { 750 return this.familyMap.keySet().toArray(new byte[0][0]); 751 } 752 return null; 753 } 754 755 /** Returns the startrow */ 756 public byte[] getStartRow() { 757 return this.startRow; 758 } 759 760 /** Returns if we should include start row when scan */ 761 public boolean includeStartRow() { 762 return includeStartRow; 763 } 764 765 /** Returns the stoprow */ 766 public byte[] getStopRow() { 767 return this.stopRow; 768 } 769 770 /** Returns if we should include stop row when scan */ 771 public boolean includeStopRow() { 772 return includeStopRow; 773 } 774 775 /** Returns the max number of versions to fetch */ 776 public int getMaxVersions() { 777 return this.maxVersions; 778 } 779 780 /** Returns maximum number of values to return for a single call to next() */ 781 public int getBatch() { 782 return this.batch; 783 } 784 785 /** Returns maximum number of values to return per row per CF */ 786 public int getMaxResultsPerColumnFamily() { 787 return this.storeLimit; 788 } 789 790 /** 791 * Method for retrieving the scan's offset per row per column family (#kvs to be skipped) 792 * @return row offset 793 */ 794 public int getRowOffsetPerColumnFamily() { 795 return this.storeOffset; 796 } 797 798 /** Returns caching the number of rows fetched when calling next on a scanner */ 799 public int getCaching() { 800 return this.caching; 801 } 802 803 /** Returns TimeRange */ 804 public TimeRange getTimeRange() { 805 return this.tr; 806 } 807 808 /** Returns RowFilter */ 809 @Override 810 public Filter getFilter() { 811 return filter; 812 } 813 814 /** Returns true is a filter has been specified, false if not */ 815 public boolean hasFilter() { 816 return filter != null; 817 } 818 819 /** 820 * Set whether blocks should be cached for this Scan. 821 * <p> 822 * This is true by default. When true, default settings of the table and family are used (this 823 * will never override caching blocks if the block cache is disabled for that family or entirely). 824 * @param cacheBlocks if false, default settings are overridden and blocks will not be cached 825 */ 826 public Scan setCacheBlocks(boolean cacheBlocks) { 827 this.cacheBlocks = cacheBlocks; 828 return this; 829 } 830 831 /** 832 * Get whether blocks should be cached for this Scan. 833 * @return true if default caching should be used, false if blocks should not be cached 834 */ 835 public boolean getCacheBlocks() { 836 return cacheBlocks; 837 } 838 839 /** 840 * Set whether this scan is a reversed one 841 * <p> 842 * This is false by default which means forward(normal) scan. 843 * @param reversed if true, scan will be backward order 844 */ 845 public Scan setReversed(boolean reversed) { 846 this.reversed = reversed; 847 return this; 848 } 849 850 /** 851 * Get whether this scan is a reversed one. 852 * @return true if backward scan, false if forward(default) scan 853 */ 854 public boolean isReversed() { 855 return reversed; 856 } 857 858 /** 859 * Setting whether the caller wants to see the partial results when server returns 860 * less-than-expected cells. It is helpful while scanning a huge row to prevent OOM at client. By 861 * default this value is false and the complete results will be assembled client side before being 862 * delivered to the caller. 863 * @see Result#mayHaveMoreCellsInRow() 864 * @see #setBatch(int) 865 */ 866 public Scan setAllowPartialResults(final boolean allowPartialResults) { 867 this.allowPartialResults = allowPartialResults; 868 return this; 869 } 870 871 /** 872 * Returns true when the constructor of this scan understands that the results they will see may 873 * only represent a partial portion of a row. The entire row would be retrieved by subsequent 874 * calls to {@link ResultScanner#next()} 875 */ 876 public boolean getAllowPartialResults() { 877 return allowPartialResults; 878 } 879 880 @Override 881 public Scan setLoadColumnFamiliesOnDemand(boolean value) { 882 super.setLoadColumnFamiliesOnDemand(value); 883 return this; 884 } 885 886 /** 887 * Compile the table and column family (i.e. schema) information into a String. Useful for parsing 888 * and aggregation by debugging, logging, and administration tools. 889 */ 890 @Override 891 public Map<String, Object> getFingerprint() { 892 Map<String, Object> map = new HashMap<>(); 893 List<String> families = new ArrayList<>(); 894 if (this.familyMap.isEmpty()) { 895 map.put("families", "ALL"); 896 return map; 897 } else { 898 map.put("families", families); 899 } 900 for (Map.Entry<byte[], NavigableSet<byte[]>> entry : this.familyMap.entrySet()) { 901 families.add(Bytes.toStringBinary(entry.getKey())); 902 } 903 return map; 904 } 905 906 /** 907 * Compile the details beyond the scope of getFingerprint (row, columns, timestamps, etc.) into a 908 * Map along with the fingerprinted information. Useful for debugging, logging, and administration 909 * tools. 910 * @param maxCols a limit on the number of columns output prior to truncation 911 */ 912 @Override 913 public Map<String, Object> toMap(int maxCols) { 914 // start with the fingerprint map and build on top of it 915 Map<String, Object> map = getFingerprint(); 916 // map from families to column list replaces fingerprint's list of families 917 Map<String, List<String>> familyColumns = new HashMap<>(); 918 map.put("families", familyColumns); 919 // add scalar information first 920 map.put("startRow", Bytes.toStringBinary(this.startRow)); 921 map.put("stopRow", Bytes.toStringBinary(this.stopRow)); 922 map.put("maxVersions", this.maxVersions); 923 map.put("batch", this.batch); 924 map.put("caching", this.caching); 925 map.put("maxResultSize", this.maxResultSize); 926 map.put("cacheBlocks", this.cacheBlocks); 927 map.put("loadColumnFamiliesOnDemand", this.loadColumnFamiliesOnDemand); 928 List<Long> timeRange = new ArrayList<>(2); 929 timeRange.add(this.tr.getMin()); 930 timeRange.add(this.tr.getMax()); 931 map.put("timeRange", timeRange); 932 int colCount = 0; 933 // iterate through affected families and list out up to maxCols columns 934 for (Map.Entry<byte[], NavigableSet<byte[]>> entry : this.familyMap.entrySet()) { 935 List<String> columns = new ArrayList<>(); 936 familyColumns.put(Bytes.toStringBinary(entry.getKey()), columns); 937 if (entry.getValue() == null) { 938 colCount++; 939 --maxCols; 940 columns.add("ALL"); 941 } else { 942 colCount += entry.getValue().size(); 943 if (maxCols <= 0) { 944 continue; 945 } 946 for (byte[] column : entry.getValue()) { 947 if (--maxCols <= 0) { 948 continue; 949 } 950 columns.add(Bytes.toStringBinary(column)); 951 } 952 } 953 } 954 map.put("totalColumns", colCount); 955 if (this.filter != null) { 956 map.put("filter", this.filter.toString()); 957 } 958 // add the id if set 959 if (getId() != null) { 960 map.put("id", getId()); 961 } 962 map.put("includeStartRow", includeStartRow); 963 map.put("includeStopRow", includeStopRow); 964 map.put("allowPartialResults", allowPartialResults); 965 map.put("storeLimit", storeLimit); 966 map.put("storeOffset", storeOffset); 967 map.put("reversed", reversed); 968 if (null != asyncPrefetch) { 969 map.put("asyncPrefetch", asyncPrefetch); 970 } 971 map.put("mvccReadPoint", mvccReadPoint); 972 map.put("limit", limit); 973 map.put("readType", readType); 974 map.put("needCursorResult", needCursorResult); 975 map.put("targetReplicaId", targetReplicaId); 976 map.put("consistency", consistency); 977 if (!colFamTimeRangeMap.isEmpty()) { 978 Map<String, List<Long>> colFamTimeRangeMapStr = colFamTimeRangeMap.entrySet().stream() 979 .collect(Collectors.toMap((e) -> Bytes.toStringBinary(e.getKey()), e -> { 980 TimeRange value = e.getValue(); 981 List<Long> rangeList = new ArrayList<>(); 982 rangeList.add(value.getMin()); 983 rangeList.add(value.getMax()); 984 return rangeList; 985 })); 986 987 map.put("colFamTimeRangeMap", colFamTimeRangeMapStr); 988 } 989 map.put("priority", getPriority()); 990 map.put("queryMetricsEnabled", queryMetricsEnabled); 991 return map; 992 } 993 994 /** 995 * Enable/disable "raw" mode for this scan. If "raw" is enabled the scan will return all delete 996 * marker and deleted rows that have not been collected, yet. This is mostly useful for Scan on 997 * column families that have KEEP_DELETED_ROWS enabled. It is an error to specify any column when 998 * "raw" is set. 999 * @param raw True/False to enable/disable "raw" mode. 1000 */ 1001 public Scan setRaw(boolean raw) { 1002 setAttribute(RAW_ATTR, Bytes.toBytes(raw)); 1003 return this; 1004 } 1005 1006 /** Returns True if this Scan is in "raw" mode. */ 1007 public boolean isRaw() { 1008 byte[] attr = getAttribute(RAW_ATTR); 1009 return attr == null ? false : Bytes.toBoolean(attr); 1010 } 1011 1012 /** 1013 * Set whether this scan is a small scan 1014 * <p> 1015 * Small scan should use pread and big scan can use seek + read seek + read is fast but can cause 1016 * two problem (1) resource contention (2) cause too much network io [89-fb] Using pread for 1017 * non-compaction read request https://issues.apache.org/jira/browse/HBASE-7266 On the other hand, 1018 * if setting it true, we would do openScanner,next,closeScanner in one RPC call. It means the 1019 * better performance for small scan. [HBASE-9488]. Generally, if the scan range is within one 1020 * data block(64KB), it could be considered as a small scan. 1021 * @param small set if that should use read type of PREAD 1022 * @deprecated since 2.0.0 and will be removed in 3.0.0. Use {@link #setLimit(int)} and 1023 * {@link #setReadType(ReadType)} instead. And for the one rpc optimization, now we 1024 * will also fetch data when openScanner, and if the number of rows reaches the limit 1025 * then we will close the scanner automatically which means we will fall back to one 1026 * rpc. 1027 * @see #setLimit(int) 1028 * @see #setReadType(ReadType) 1029 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17045">HBASE-17045</a> 1030 */ 1031 @Deprecated 1032 public Scan setSmall(boolean small) { 1033 this.small = small; 1034 if (small) { 1035 this.readType = ReadType.PREAD; 1036 } 1037 return this; 1038 } 1039 1040 /** 1041 * Get whether this scan is a small scan 1042 * @return true if small scan 1043 * @deprecated since 2.0.0 and will be removed in 3.0.0. See the comment of 1044 * {@link #setSmall(boolean)} 1045 * @see <a href="https://issues.apache.org/jira/browse/HBASE-17045">HBASE-17045</a> 1046 */ 1047 @Deprecated 1048 public boolean isSmall() { 1049 return small; 1050 } 1051 1052 @Override 1053 public Scan setAttribute(String name, byte[] value) { 1054 super.setAttribute(name, value); 1055 return this; 1056 } 1057 1058 @Override 1059 public Scan setId(String id) { 1060 super.setId(id); 1061 return this; 1062 } 1063 1064 @Override 1065 public Scan setAuthorizations(Authorizations authorizations) { 1066 super.setAuthorizations(authorizations); 1067 return this; 1068 } 1069 1070 @Override 1071 public Scan setACL(Map<String, Permission> perms) { 1072 super.setACL(perms); 1073 return this; 1074 } 1075 1076 @Override 1077 public Scan setACL(String user, Permission perms) { 1078 super.setACL(user, perms); 1079 return this; 1080 } 1081 1082 @Override 1083 public Scan setConsistency(Consistency consistency) { 1084 super.setConsistency(consistency); 1085 return this; 1086 } 1087 1088 @Override 1089 public Scan setReplicaId(int Id) { 1090 super.setReplicaId(Id); 1091 return this; 1092 } 1093 1094 @Override 1095 public Scan setIsolationLevel(IsolationLevel level) { 1096 super.setIsolationLevel(level); 1097 return this; 1098 } 1099 1100 @Override 1101 public Scan setPriority(int priority) { 1102 super.setPriority(priority); 1103 return this; 1104 } 1105 1106 /** 1107 * Enable collection of {@link ScanMetrics}. For advanced users. While disabling scan metrics, 1108 * will also disable region level scan metrics. 1109 * @param enabled Set to true to enable accumulating scan metrics 1110 */ 1111 public Scan setScanMetricsEnabled(final boolean enabled) { 1112 setAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE, Bytes.toBytes(Boolean.valueOf(enabled))); 1113 if (!enabled) { 1114 setEnableScanMetricsByRegion(false); 1115 } 1116 return this; 1117 } 1118 1119 /** Returns True if collection of scan metrics is enabled. For advanced users. */ 1120 public boolean isScanMetricsEnabled() { 1121 byte[] attr = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE); 1122 return attr == null ? false : Bytes.toBoolean(attr); 1123 } 1124 1125 /** 1126 * @return Metrics on this Scan, if metrics were enabled. 1127 * @see #setScanMetricsEnabled(boolean) 1128 * @deprecated Use {@link ResultScanner#getScanMetrics()} instead. And notice that, please do not 1129 * use this method and {@link ResultScanner#getScanMetrics()} together, the metrics 1130 * will be messed up. 1131 */ 1132 @Deprecated 1133 public ScanMetrics getScanMetrics() { 1134 byte[] bytes = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_DATA); 1135 if (bytes == null) return null; 1136 return ProtobufUtil.toScanMetrics(bytes); 1137 } 1138 1139 public Boolean isAsyncPrefetch() { 1140 return asyncPrefetch; 1141 } 1142 1143 public Scan setAsyncPrefetch(boolean asyncPrefetch) { 1144 this.asyncPrefetch = asyncPrefetch; 1145 return this; 1146 } 1147 1148 /** Returns the limit of rows for this scan */ 1149 public int getLimit() { 1150 return limit; 1151 } 1152 1153 /** 1154 * Set the limit of rows for this scan. We will terminate the scan if the number of returned rows 1155 * reaches this value. 1156 * <p> 1157 * This condition will be tested at last, after all other conditions such as stopRow, filter, etc. 1158 * @param limit the limit of rows for this scan 1159 */ 1160 public Scan setLimit(int limit) { 1161 this.limit = limit; 1162 return this; 1163 } 1164 1165 /** 1166 * Call this when you only want to get one row. It will set {@code limit} to {@code 1}, and also 1167 * set {@code readType} to {@link ReadType#PREAD}. 1168 */ 1169 public Scan setOneRowLimit() { 1170 return setLimit(1).setReadType(ReadType.PREAD); 1171 } 1172 1173 @InterfaceAudience.Public 1174 public enum ReadType { 1175 DEFAULT, 1176 STREAM, 1177 PREAD 1178 } 1179 1180 /** Returns the read type for this scan */ 1181 public ReadType getReadType() { 1182 return readType; 1183 } 1184 1185 /** 1186 * Set the read type for this scan. 1187 * <p> 1188 * Notice that we may choose to use pread even if you specific {@link ReadType#STREAM} here. For 1189 * example, we will always use pread if this is a get scan. 1190 */ 1191 public Scan setReadType(ReadType readType) { 1192 this.readType = readType; 1193 return this; 1194 } 1195 1196 /** 1197 * Get the mvcc read point used to open a scanner. 1198 */ 1199 long getMvccReadPoint() { 1200 return mvccReadPoint; 1201 } 1202 1203 /** 1204 * Set the mvcc read point used to open a scanner. 1205 */ 1206 Scan setMvccReadPoint(long mvccReadPoint) { 1207 this.mvccReadPoint = mvccReadPoint; 1208 return this; 1209 } 1210 1211 /** 1212 * Set the mvcc read point to -1 which means do not use it. 1213 */ 1214 Scan resetMvccReadPoint() { 1215 return setMvccReadPoint(-1L); 1216 } 1217 1218 /** 1219 * When the server is slow or we scan a table with many deleted data or we use a sparse filter, 1220 * the server will response heartbeat to prevent timeout. However the scanner will return a Result 1221 * only when client can do it. So if there are many heartbeats, the blocking time on 1222 * ResultScanner#next() may be very long, which is not friendly to online services. Set this to 1223 * true then you can get a special Result whose #isCursor() returns true and is not contains any 1224 * real data. It only tells you where the server has scanned. You can call next to continue 1225 * scanning or open a new scanner with this row key as start row whenever you want. Users can get 1226 * a cursor when and only when there is a response from the server but we can not return a Result 1227 * to users, for example, this response is a heartbeat or there are partial cells but users do not 1228 * allow partial result. Now the cursor is in row level which means the special Result will only 1229 * contains a row key. {@link Result#isCursor()} {@link Result#getCursor()} {@link Cursor} 1230 */ 1231 public Scan setNeedCursorResult(boolean needCursorResult) { 1232 this.needCursorResult = needCursorResult; 1233 return this; 1234 } 1235 1236 public boolean isNeedCursorResult() { 1237 return needCursorResult; 1238 } 1239 1240 /** 1241 * Create a new Scan with a cursor. It only set the position information like start row key. The 1242 * others (like cfs, stop row, limit) should still be filled in by the user. 1243 * {@link Result#isCursor()} {@link Result#getCursor()} {@link Cursor} 1244 */ 1245 public static Scan createScanFromCursor(Cursor cursor) { 1246 return new Scan().withStartRow(cursor.getRow()); 1247 } 1248 1249 /** 1250 * Enables region level scan metrics. If scan metrics are disabled then first enables scan metrics 1251 * followed by region level scan metrics. 1252 * @param enable Set to true to enable region level scan metrics. 1253 */ 1254 public Scan setEnableScanMetricsByRegion(final boolean enable) { 1255 if (enable) { 1256 setScanMetricsEnabled(true); 1257 } 1258 setAttribute(Scan.SCAN_ATTRIBUTES_METRICS_BY_REGION_ENABLE, Bytes.toBytes(enable)); 1259 return this; 1260 } 1261 1262 public boolean isScanMetricsByRegionEnabled() { 1263 byte[] attr = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_BY_REGION_ENABLE); 1264 return attr != null && Bytes.toBoolean(attr); 1265 } 1266}