1 /*
2 *
3 * Licensed to the Apache Software Foundation (ASF) under one
4 * or more contributor license agreements. See the NOTICE file
5 * distributed with this work for additional information
6 * regarding copyright ownership. The ASF licenses this file
7 * to you under the Apache License, Version 2.0 (the
8 * "License"); you may not use this file except in compliance
9 * with the License. You may obtain a copy of the License at
10 *
11 * http://www.apache.org/licenses/LICENSE-2.0
12 *
13 * Unless required by applicable law or agreed to in writing, software
14 * distributed under the License is distributed on an "AS IS" BASIS,
15 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16 * See the License for the specific language governing permissions and
17 * limitations under the License.
18 */
19
20 package org.apache.hadoop.hbase.client;
21
22 import java.io.IOException;
23 import java.util.ArrayList;
24 import java.util.Arrays;
25 import java.util.HashMap;
26 import java.util.List;
27 import java.util.Map;
28 import java.util.NavigableSet;
29 import java.util.TreeMap;
30 import java.util.TreeSet;
31
32 import org.apache.commons.logging.Log;
33 import org.apache.commons.logging.LogFactory;
34 import org.apache.hadoop.hbase.HConstants;
35 import org.apache.hadoop.hbase.classification.InterfaceAudience;
36 import org.apache.hadoop.hbase.classification.InterfaceStability;
37 import org.apache.hadoop.hbase.client.metrics.ScanMetrics;
38 import org.apache.hadoop.hbase.filter.Filter;
39 import org.apache.hadoop.hbase.filter.IncompatibleFilterException;
40 import org.apache.hadoop.hbase.io.TimeRange;
41 import org.apache.hadoop.hbase.protobuf.ProtobufUtil;
42 import org.apache.hadoop.hbase.security.access.Permission;
43 import org.apache.hadoop.hbase.security.visibility.Authorizations;
44 import org.apache.hadoop.hbase.util.Bytes;
45
46 /**
47 * Used to perform Scan operations.
48 * <p>
49 * All operations are identical to {@link Get} with the exception of
50 * instantiation. Rather than specifying a single row, an optional startRow
51 * and stopRow may be defined. If rows are not specified, the Scanner will
52 * iterate over all rows.
53 * <p>
54 * To get all columns from all rows of a Table, create an instance with no constraints; use the
55 * {@link #Scan()} constructor. To constrain the scan to specific column families,
56 * call {@link #addFamily(byte[]) addFamily} for each family to retrieve on your Scan instance.
57 * <p>
58 * To get specific columns, call {@link #addColumn(byte[], byte[]) addColumn}
59 * for each column to retrieve.
60 * <p>
61 * To only retrieve columns within a specific range of version timestamps,
62 * call {@link #setTimeRange(long, long) setTimeRange}.
63 * <p>
64 * To only retrieve columns with a specific timestamp, call
65 * {@link #setTimeStamp(long) setTimestamp}.
66 * <p>
67 * To limit the number of versions of each column to be returned, call
68 * {@link #setMaxVersions(int) setMaxVersions}.
69 * <p>
70 * To limit the maximum number of values returned for each call to next(),
71 * call {@link #setBatch(int) setBatch}.
72 * <p>
73 * To add a filter, call {@link #setFilter(org.apache.hadoop.hbase.filter.Filter) setFilter}.
74 * <p>
75 * Expert: To explicitly disable server-side block caching for this scan,
76 * execute {@link #setCacheBlocks(boolean)}.
77 * <p><em>Note:</em> Usage alters Scan instances. Internally, attributes are updated as the Scan
78 * runs and if enabled, metrics accumulate in the Scan instance. Be aware this is the case when
79 * you go to clone a Scan instance or if you go to reuse a created Scan instance; safer is create
80 * a Scan instance per usage.
81 */
82 @InterfaceAudience.Public
83 @InterfaceStability.Stable
84 public class Scan extends Query {
85 private static final Log LOG = LogFactory.getLog(Scan.class);
86
87 private static final String RAW_ATTR = "_raw_";
88
89 private byte [] startRow = HConstants.EMPTY_START_ROW;
90 private byte [] stopRow = HConstants.EMPTY_END_ROW;
91 private int maxVersions = 1;
92 private int batch = -1;
93
94 /**
95 * Partial {@link Result}s are {@link Result}s must be combined to form a complete {@link Result}.
96 * The {@link Result}s had to be returned in fragments (i.e. as partials) because the size of the
97 * cells in the row exceeded max result size on the server. Typically partial results will be
98 * combined client side into complete results before being delivered to the caller. However, if
99 * this flag is set, the caller is indicating that they do not mind seeing partial results (i.e.
100 * they understand that the results returned from the Scanner may only represent part of a
101 * particular row). In such a case, any attempt to combine the partials into a complete result on
102 * the client side will be skipped, and the caller will be able to see the exact results returned
103 * from the server.
104 */
105 private boolean allowPartialResults = false;
106
107 private int storeLimit = -1;
108 private int storeOffset = 0;
109 private boolean getScan;
110
111 /**
112 * @deprecated since 1.0.0. Use {@link #setScanMetricsEnabled(boolean)}
113 */
114 // Make private or remove.
115 @Deprecated
116 static public final String SCAN_ATTRIBUTES_METRICS_ENABLE = "scan.attributes.metrics.enable";
117
118 /**
119 * Use {@link #getScanMetrics()}
120 */
121 // Make this private or remove.
122 @Deprecated
123 static public final String SCAN_ATTRIBUTES_METRICS_DATA = "scan.attributes.metrics.data";
124
125 // If an application wants to use multiple scans over different tables each scan must
126 // define this attribute with the appropriate table name by calling
127 // scan.setAttribute(Scan.SCAN_ATTRIBUTES_TABLE_NAME, Bytes.toBytes(tableName))
128 static public final String SCAN_ATTRIBUTES_TABLE_NAME = "scan.attributes.table.name";
129
130 /*
131 * -1 means no caching
132 */
133 private int caching = -1;
134 private long maxResultSize = -1;
135 private boolean cacheBlocks = true;
136 private boolean reversed = false;
137 private TimeRange tr = new TimeRange();
138 private Map<byte [], NavigableSet<byte []>> familyMap =
139 new TreeMap<byte [], NavigableSet<byte []>>(Bytes.BYTES_COMPARATOR);
140 private Boolean loadColumnFamiliesOnDemand = null;
141 private Boolean asyncPrefetch = null;
142
143 /**
144 * Parameter name for client scanner sync/async prefetch toggle.
145 * When using async scanner, prefetching data from the server is done at the background.
146 * The parameter currently won't have any effect in the case that the user has set
147 * Scan#setSmall or Scan#setReversed
148 */
149 public static final String HBASE_CLIENT_SCANNER_ASYNC_PREFETCH =
150 "hbase.client.scanner.async.prefetch";
151
152 /**
153 * Default value of {@link #HBASE_CLIENT_SCANNER_ASYNC_PREFETCH}.
154 */
155 public static final boolean DEFAULT_HBASE_CLIENT_SCANNER_ASYNC_PREFETCH = false;
156
157 /**
158 * Set it true for small scan to get better performance
159 *
160 * Small scan should use pread and big scan can use seek + read
161 *
162 * seek + read is fast but can cause two problem (1) resource contention (2)
163 * cause too much network io
164 *
165 * [89-fb] Using pread for non-compaction read request
166 * https://issues.apache.org/jira/browse/HBASE-7266
167 *
168 * On the other hand, if setting it true, we would do
169 * openScanner,next,closeScanner in one RPC call. It means the better
170 * performance for small scan. [HBASE-9488].
171 *
172 * Generally, if the scan range is within one data block(64KB), it could be
173 * considered as a small scan.
174 */
175 private boolean small = false;
176
177 /**
178 * Create a Scan operation across all rows.
179 */
180 public Scan() {}
181
182 public Scan(byte [] startRow, Filter filter) {
183 this(startRow);
184 this.filter = filter;
185 }
186
187 /**
188 * Create a Scan operation starting at the specified row.
189 * <p>
190 * If the specified row does not exist, the Scanner will start from the
191 * next closest row after the specified row.
192 * @param startRow row to start scanner at or after
193 */
194 public Scan(byte [] startRow) {
195 this.startRow = startRow;
196 }
197
198 /**
199 * Create a Scan operation for the range of rows specified.
200 * @param startRow row to start scanner at or after (inclusive)
201 * @param stopRow row to stop scanner before (exclusive)
202 */
203 public Scan(byte [] startRow, byte [] stopRow) {
204 this.startRow = startRow;
205 this.stopRow = stopRow;
206 //if the startRow and stopRow both are empty, it is not a Get
207 this.getScan = isStartRowAndEqualsStopRow();
208 }
209
210 /**
211 * Creates a new instance of this class while copying all values.
212 *
213 * @param scan The scan instance to copy from.
214 * @throws IOException When copying the values fails.
215 */
216 public Scan(Scan scan) throws IOException {
217 startRow = scan.getStartRow();
218 stopRow = scan.getStopRow();
219 maxVersions = scan.getMaxVersions();
220 batch = scan.getBatch();
221 storeLimit = scan.getMaxResultsPerColumnFamily();
222 storeOffset = scan.getRowOffsetPerColumnFamily();
223 caching = scan.getCaching();
224 maxResultSize = scan.getMaxResultSize();
225 cacheBlocks = scan.getCacheBlocks();
226 getScan = scan.isGetScan();
227 filter = scan.getFilter(); // clone?
228 loadColumnFamiliesOnDemand = scan.getLoadColumnFamiliesOnDemandValue();
229 consistency = scan.getConsistency();
230 reversed = scan.isReversed();
231 asyncPrefetch = scan.isAsyncPrefetch();
232 small = scan.isSmall();
233 allowPartialResults = scan.getAllowPartialResults();
234 TimeRange ctr = scan.getTimeRange();
235 tr = new TimeRange(ctr.getMin(), ctr.getMax());
236 Map<byte[], NavigableSet<byte[]>> fams = scan.getFamilyMap();
237 for (Map.Entry<byte[],NavigableSet<byte[]>> entry : fams.entrySet()) {
238 byte [] fam = entry.getKey();
239 NavigableSet<byte[]> cols = entry.getValue();
240 if (cols != null && cols.size() > 0) {
241 for (byte[] col : cols) {
242 addColumn(fam, col);
243 }
244 } else {
245 addFamily(fam);
246 }
247 }
248 for (Map.Entry<String, byte[]> attr : scan.getAttributesMap().entrySet()) {
249 setAttribute(attr.getKey(), attr.getValue());
250 }
251 }
252
253 /**
254 * Builds a scan object with the same specs as get.
255 * @param get get to model scan after
256 */
257 public Scan(Get get) {
258 this.startRow = get.getRow();
259 this.stopRow = get.getRow();
260 this.filter = get.getFilter();
261 this.cacheBlocks = get.getCacheBlocks();
262 this.maxVersions = get.getMaxVersions();
263 this.storeLimit = get.getMaxResultsPerColumnFamily();
264 this.storeOffset = get.getRowOffsetPerColumnFamily();
265 this.tr = get.getTimeRange();
266 this.familyMap = get.getFamilyMap();
267 this.getScan = true;
268 this.asyncPrefetch = false;
269 this.consistency = get.getConsistency();
270 for (Map.Entry<String, byte[]> attr : get.getAttributesMap().entrySet()) {
271 setAttribute(attr.getKey(), attr.getValue());
272 }
273 }
274
275 public boolean isGetScan() {
276 return this.getScan || isStartRowAndEqualsStopRow();
277 }
278
279 private boolean isStartRowAndEqualsStopRow() {
280 return this.startRow != null && this.startRow.length > 0 &&
281 Bytes.equals(this.startRow, this.stopRow);
282 }
283 /**
284 * Get all columns from the specified family.
285 * <p>
286 * Overrides previous calls to addColumn for this family.
287 * @param family family name
288 * @return this
289 */
290 public Scan addFamily(byte [] family) {
291 familyMap.remove(family);
292 familyMap.put(family, null);
293 return this;
294 }
295
296 /**
297 * Get the column from the specified family with the specified qualifier.
298 * <p>
299 * Overrides previous calls to addFamily for this family.
300 * @param family family name
301 * @param qualifier column qualifier
302 * @return this
303 */
304 public Scan addColumn(byte [] family, byte [] qualifier) {
305 NavigableSet<byte []> set = familyMap.get(family);
306 if(set == null) {
307 set = new TreeSet<byte []>(Bytes.BYTES_COMPARATOR);
308 }
309 if (qualifier == null) {
310 qualifier = HConstants.EMPTY_BYTE_ARRAY;
311 }
312 set.add(qualifier);
313 familyMap.put(family, set);
314 return this;
315 }
316
317 /**
318 * Get versions of columns only within the specified timestamp range,
319 * [minStamp, maxStamp). Note, default maximum versions to return is 1. If
320 * your time range spans more than one version and you want all versions
321 * returned, up the number of versions beyond the default.
322 * @param minStamp minimum timestamp value, inclusive
323 * @param maxStamp maximum timestamp value, exclusive
324 * @throws IOException if invalid time range
325 * @see #setMaxVersions()
326 * @see #setMaxVersions(int)
327 * @return this
328 */
329 public Scan setTimeRange(long minStamp, long maxStamp)
330 throws IOException {
331 tr = new TimeRange(minStamp, maxStamp);
332 return this;
333 }
334
335 /**
336 * Get versions of columns with the specified timestamp. Note, default maximum
337 * versions to return is 1. If your time range spans more than one version
338 * and you want all versions returned, up the number of versions beyond the
339 * defaut.
340 * @param timestamp version timestamp
341 * @see #setMaxVersions()
342 * @see #setMaxVersions(int)
343 * @return this
344 */
345 public Scan setTimeStamp(long timestamp)
346 throws IOException {
347 try {
348 tr = new TimeRange(timestamp, timestamp+1);
349 } catch(IOException e) {
350 // This should never happen, unless integer overflow or something extremely wrong...
351 LOG.error("TimeRange failed, likely caused by integer overflow. ", e);
352 throw e;
353 }
354 return this;
355 }
356
357 /**
358 * Set the start row of the scan.
359 * <p>
360 * If the specified row does not exist, the Scanner will start from the
361 * next closest row after the specified row.
362 * @param startRow row to start scanner at or after
363 * @return this
364 */
365 public Scan setStartRow(byte [] startRow) {
366 this.startRow = startRow;
367 return this;
368 }
369
370 /**
371 * Set the stop row of the scan.
372 * @param stopRow row to end at (exclusive)
373 * <p>
374 * The scan will include rows that are lexicographically less than
375 * the provided stopRow.
376 * <p><b>Note:</b> When doing a filter for a rowKey <u>Prefix</u>
377 * use {@link #setRowPrefixFilter(byte[])}.
378 * The 'trailing 0' will not yield the desired result.</p>
379 * @return this
380 */
381 public Scan setStopRow(byte [] stopRow) {
382 this.stopRow = stopRow;
383 return this;
384 }
385
386 /**
387 * <p>Set a filter (using stopRow and startRow) so the result set only contains rows where the
388 * rowKey starts with the specified prefix.</p>
389 * <p>This is a utility method that converts the desired rowPrefix into the appropriate values
390 * for the startRow and stopRow to achieve the desired result.</p>
391 * <p>This can safely be used in combination with setFilter.</p>
392 * <p><b>NOTE: Doing a {@link #setStartRow(byte[])} and/or {@link #setStopRow(byte[])}
393 * after this method will yield undefined results.</b></p>
394 * @param rowPrefix the prefix all rows must start with. (Set <i>null</i> to remove the filter.)
395 * @return this
396 */
397 public Scan setRowPrefixFilter(byte[] rowPrefix) {
398 if (rowPrefix == null) {
399 setStartRow(HConstants.EMPTY_START_ROW);
400 setStopRow(HConstants.EMPTY_END_ROW);
401 } else {
402 this.setStartRow(rowPrefix);
403 this.setStopRow(calculateTheClosestNextRowKeyForPrefix(rowPrefix));
404 }
405 return this;
406 }
407
408 /**
409 * <p>When scanning for a prefix the scan should stop immediately after the the last row that
410 * has the specified prefix. This method calculates the closest next rowKey immediately following
411 * the given rowKeyPrefix.</p>
412 * <p><b>IMPORTANT: This converts a rowKey<u>Prefix</u> into a rowKey</b>.</p>
413 * <p>If the prefix is an 'ASCII' string put into a byte[] then this is easy because you can
414 * simply increment the last byte of the array.
415 * But if your application uses real binary rowids you may run into the scenario that your
416 * prefix is something like:</p>
417 * <b>{ 0x12, 0x23, 0xFF, 0xFF }</b><br/>
418 * Then this stopRow needs to be fed into the actual scan<br/>
419 * <b>{ 0x12, 0x24 }</b> (Notice that it is shorter now)<br/>
420 * This method calculates the correct stop row value for this usecase.
421 *
422 * @param rowKeyPrefix the rowKey<u>Prefix</u>.
423 * @return the closest next rowKey immediately following the given rowKeyPrefix.
424 */
425 private byte[] calculateTheClosestNextRowKeyForPrefix(byte[] rowKeyPrefix) {
426 // Essentially we are treating it like an 'unsigned very very long' and doing +1 manually.
427 // Search for the place where the trailing 0xFFs start
428 int offset = rowKeyPrefix.length;
429 while (offset > 0) {
430 if (rowKeyPrefix[offset - 1] != (byte) 0xFF) {
431 break;
432 }
433 offset--;
434 }
435
436 if (offset == 0) {
437 // We got an 0xFFFF... (only FFs) stopRow value which is
438 // the last possible prefix before the end of the table.
439 // So set it to stop at the 'end of the table'
440 return HConstants.EMPTY_END_ROW;
441 }
442
443 // Copy the right length of the original
444 byte[] newStopRow = Arrays.copyOfRange(rowKeyPrefix, 0, offset);
445 // And increment the last one
446 newStopRow[newStopRow.length - 1]++;
447 return newStopRow;
448 }
449
450 /**
451 * Get all available versions.
452 * @return this
453 */
454 public Scan setMaxVersions() {
455 this.maxVersions = Integer.MAX_VALUE;
456 return this;
457 }
458
459 /**
460 * Get up to the specified number of versions of each column.
461 * @param maxVersions maximum versions for each column
462 * @return this
463 */
464 public Scan setMaxVersions(int maxVersions) {
465 this.maxVersions = maxVersions;
466 return this;
467 }
468
469 /**
470 * Set the maximum number of values to return for each call to next()
471 * @param batch the maximum number of values
472 */
473 public Scan setBatch(int batch) {
474 if (this.hasFilter() && this.filter.hasFilterRow()) {
475 throw new IncompatibleFilterException(
476 "Cannot set batch on a scan using a filter" +
477 " that returns true for filter.hasFilterRow");
478 }
479 this.batch = batch;
480 return this;
481 }
482
483 /**
484 * Set the maximum number of values to return per row per Column Family
485 * @param limit the maximum number of values returned / row / CF
486 */
487 public Scan setMaxResultsPerColumnFamily(int limit) {
488 this.storeLimit = limit;
489 return this;
490 }
491
492 /**
493 * Set offset for the row per Column Family.
494 * @param offset is the number of kvs that will be skipped.
495 */
496 public Scan setRowOffsetPerColumnFamily(int offset) {
497 this.storeOffset = offset;
498 return this;
499 }
500
501 /**
502 * Set the number of rows for caching that will be passed to scanners.
503 * If not set, the Configuration setting {@link HConstants#HBASE_CLIENT_SCANNER_CACHING} will
504 * apply.
505 * Higher caching values will enable faster scanners but will use more memory.
506 * @param caching the number of rows for caching
507 */
508 public Scan setCaching(int caching) {
509 this.caching = caching;
510 return this;
511 }
512
513 /**
514 * @return the maximum result size in bytes. See {@link #setMaxResultSize(long)}
515 */
516 public long getMaxResultSize() {
517 return maxResultSize;
518 }
519
520 /**
521 * Set the maximum result size. The default is -1; this means that no specific
522 * maximum result size will be set for this scan, and the global configured
523 * value will be used instead. (Defaults to unlimited).
524 *
525 * @param maxResultSize The maximum result size in bytes.
526 */
527 public Scan setMaxResultSize(long maxResultSize) {
528 this.maxResultSize = maxResultSize;
529 return this;
530 }
531
532 @Override
533 public Scan setFilter(Filter filter) {
534 super.setFilter(filter);
535 return this;
536 }
537
538 /**
539 * Setting the familyMap
540 * @param familyMap map of family to qualifier
541 * @return this
542 */
543 public Scan setFamilyMap(Map<byte [], NavigableSet<byte []>> familyMap) {
544 this.familyMap = familyMap;
545 return this;
546 }
547
548 /**
549 * Getting the familyMap
550 * @return familyMap
551 */
552 public Map<byte [], NavigableSet<byte []>> getFamilyMap() {
553 return this.familyMap;
554 }
555
556 /**
557 * @return the number of families in familyMap
558 */
559 public int numFamilies() {
560 if(hasFamilies()) {
561 return this.familyMap.size();
562 }
563 return 0;
564 }
565
566 /**
567 * @return true if familyMap is non empty, false otherwise
568 */
569 public boolean hasFamilies() {
570 return !this.familyMap.isEmpty();
571 }
572
573 /**
574 * @return the keys of the familyMap
575 */
576 public byte[][] getFamilies() {
577 if(hasFamilies()) {
578 return this.familyMap.keySet().toArray(new byte[0][0]);
579 }
580 return null;
581 }
582
583 /**
584 * @return the startrow
585 */
586 public byte [] getStartRow() {
587 return this.startRow;
588 }
589
590 /**
591 * @return the stoprow
592 */
593 public byte [] getStopRow() {
594 return this.stopRow;
595 }
596
597 /**
598 * @return the max number of versions to fetch
599 */
600 public int getMaxVersions() {
601 return this.maxVersions;
602 }
603
604 /**
605 * @return maximum number of values to return for a single call to next()
606 */
607 public int getBatch() {
608 return this.batch;
609 }
610
611 /**
612 * @return maximum number of values to return per row per CF
613 */
614 public int getMaxResultsPerColumnFamily() {
615 return this.storeLimit;
616 }
617
618 /**
619 * Method for retrieving the scan's offset per row per column
620 * family (#kvs to be skipped)
621 * @return row offset
622 */
623 public int getRowOffsetPerColumnFamily() {
624 return this.storeOffset;
625 }
626
627 /**
628 * @return caching the number of rows fetched when calling next on a scanner
629 */
630 public int getCaching() {
631 return this.caching;
632 }
633
634 /**
635 * @return TimeRange
636 */
637 public TimeRange getTimeRange() {
638 return this.tr;
639 }
640
641 /**
642 * @return RowFilter
643 */
644 @Override
645 public Filter getFilter() {
646 return filter;
647 }
648
649 /**
650 * @return true is a filter has been specified, false if not
651 */
652 public boolean hasFilter() {
653 return filter != null;
654 }
655
656 /**
657 * Set whether blocks should be cached for this Scan.
658 * <p>
659 * This is true by default. When true, default settings of the table and
660 * family are used (this will never override caching blocks if the block
661 * cache is disabled for that family or entirely).
662 *
663 * @param cacheBlocks if false, default settings are overridden and blocks
664 * will not be cached
665 */
666 public Scan setCacheBlocks(boolean cacheBlocks) {
667 this.cacheBlocks = cacheBlocks;
668 return this;
669 }
670
671 /**
672 * Get whether blocks should be cached for this Scan.
673 * @return true if default caching should be used, false if blocks should not
674 * be cached
675 */
676 public boolean getCacheBlocks() {
677 return cacheBlocks;
678 }
679
680 /**
681 * Set whether this scan is a reversed one
682 * <p>
683 * This is false by default which means forward(normal) scan.
684 *
685 * @param reversed if true, scan will be backward order
686 * @return this
687 */
688 public Scan setReversed(boolean reversed) {
689 this.reversed = reversed;
690 return this;
691 }
692
693 /**
694 * Get whether this scan is a reversed one.
695 * @return true if backward scan, false if forward(default) scan
696 */
697 public boolean isReversed() {
698 return reversed;
699 }
700
701 /**
702 * Setting whether the caller wants to see the partial results that may be returned from the
703 * server. By default this value is false and the complete results will be assembled client side
704 * before being delivered to the caller.
705 * @param allowPartialResults
706 * @return this
707 */
708 public Scan setAllowPartialResults(final boolean allowPartialResults) {
709 this.allowPartialResults = allowPartialResults;
710 return this;
711 }
712
713 /**
714 * @return true when the constructor of this scan understands that the results they will see may
715 * only represent a partial portion of a row. The entire row would be retrieved by
716 * subsequent calls to {@link ResultScanner#next()}
717 */
718 public boolean getAllowPartialResults() {
719 return allowPartialResults;
720 }
721
722 /**
723 * Set the value indicating whether loading CFs on demand should be allowed (cluster
724 * default is false). On-demand CF loading doesn't load column families until necessary, e.g.
725 * if you filter on one column, the other column family data will be loaded only for the rows
726 * that are included in result, not all rows like in normal case.
727 * With column-specific filters, like SingleColumnValueFilter w/filterIfMissing == true,
728 * this can deliver huge perf gains when there's a cf with lots of data; however, it can
729 * also lead to some inconsistent results, as follows:
730 * - if someone does a concurrent update to both column families in question you may get a row
731 * that never existed, e.g. for { rowKey = 5, { cat_videos => 1 }, { video => "my cat" } }
732 * someone puts rowKey 5 with { cat_videos => 0 }, { video => "my dog" }, concurrent scan
733 * filtering on "cat_videos == 1" can get { rowKey = 5, { cat_videos => 1 },
734 * { video => "my dog" } }.
735 * - if there's a concurrent split and you have more than 2 column families, some rows may be
736 * missing some column families.
737 */
738 public Scan setLoadColumnFamiliesOnDemand(boolean value) {
739 this.loadColumnFamiliesOnDemand = value;
740 return this;
741 }
742
743 /**
744 * Get the raw loadColumnFamiliesOnDemand setting; if it's not set, can be null.
745 */
746 public Boolean getLoadColumnFamiliesOnDemandValue() {
747 return this.loadColumnFamiliesOnDemand;
748 }
749
750 /**
751 * Get the logical value indicating whether on-demand CF loading should be allowed.
752 */
753 public boolean doLoadColumnFamiliesOnDemand() {
754 return (this.loadColumnFamiliesOnDemand != null)
755 && this.loadColumnFamiliesOnDemand.booleanValue();
756 }
757
758 /**
759 * Compile the table and column family (i.e. schema) information
760 * into a String. Useful for parsing and aggregation by debugging,
761 * logging, and administration tools.
762 * @return Map
763 */
764 @Override
765 public Map<String, Object> getFingerprint() {
766 Map<String, Object> map = new HashMap<String, Object>();
767 List<String> families = new ArrayList<String>();
768 if(this.familyMap.size() == 0) {
769 map.put("families", "ALL");
770 return map;
771 } else {
772 map.put("families", families);
773 }
774 for (Map.Entry<byte [], NavigableSet<byte[]>> entry :
775 this.familyMap.entrySet()) {
776 families.add(Bytes.toStringBinary(entry.getKey()));
777 }
778 return map;
779 }
780
781 /**
782 * Compile the details beyond the scope of getFingerprint (row, columns,
783 * timestamps, etc.) into a Map along with the fingerprinted information.
784 * Useful for debugging, logging, and administration tools.
785 * @param maxCols a limit on the number of columns output prior to truncation
786 * @return Map
787 */
788 @Override
789 public Map<String, Object> toMap(int maxCols) {
790 // start with the fingerpring map and build on top of it
791 Map<String, Object> map = getFingerprint();
792 // map from families to column list replaces fingerprint's list of families
793 Map<String, List<String>> familyColumns =
794 new HashMap<String, List<String>>();
795 map.put("families", familyColumns);
796 // add scalar information first
797 map.put("startRow", Bytes.toStringBinary(this.startRow));
798 map.put("stopRow", Bytes.toStringBinary(this.stopRow));
799 map.put("maxVersions", this.maxVersions);
800 map.put("batch", this.batch);
801 map.put("caching", this.caching);
802 map.put("maxResultSize", this.maxResultSize);
803 map.put("cacheBlocks", this.cacheBlocks);
804 map.put("loadColumnFamiliesOnDemand", this.loadColumnFamiliesOnDemand);
805 List<Long> timeRange = new ArrayList<Long>();
806 timeRange.add(this.tr.getMin());
807 timeRange.add(this.tr.getMax());
808 map.put("timeRange", timeRange);
809 int colCount = 0;
810 // iterate through affected families and list out up to maxCols columns
811 for (Map.Entry<byte [], NavigableSet<byte[]>> entry :
812 this.familyMap.entrySet()) {
813 List<String> columns = new ArrayList<String>();
814 familyColumns.put(Bytes.toStringBinary(entry.getKey()), columns);
815 if(entry.getValue() == null) {
816 colCount++;
817 --maxCols;
818 columns.add("ALL");
819 } else {
820 colCount += entry.getValue().size();
821 if (maxCols <= 0) {
822 continue;
823 }
824 for (byte [] column : entry.getValue()) {
825 if (--maxCols <= 0) {
826 continue;
827 }
828 columns.add(Bytes.toStringBinary(column));
829 }
830 }
831 }
832 map.put("totalColumns", colCount);
833 if (this.filter != null) {
834 map.put("filter", this.filter.toString());
835 }
836 // add the id if set
837 if (getId() != null) {
838 map.put("id", getId());
839 }
840 return map;
841 }
842
843 /**
844 * Enable/disable "raw" mode for this scan.
845 * If "raw" is enabled the scan will return all
846 * delete marker and deleted rows that have not
847 * been collected, yet.
848 * This is mostly useful for Scan on column families
849 * that have KEEP_DELETED_ROWS enabled.
850 * It is an error to specify any column when "raw" is set.
851 * @param raw True/False to enable/disable "raw" mode.
852 */
853 public Scan setRaw(boolean raw) {
854 setAttribute(RAW_ATTR, Bytes.toBytes(raw));
855 return this;
856 }
857
858 /**
859 * @return True if this Scan is in "raw" mode.
860 */
861 public boolean isRaw() {
862 byte[] attr = getAttribute(RAW_ATTR);
863 return attr == null ? false : Bytes.toBoolean(attr);
864 }
865
866
867
868 /**
869 * Set whether this scan is a small scan
870 * <p>
871 * Small scan should use pread and big scan can use seek + read
872 *
873 * seek + read is fast but can cause two problem (1) resource contention (2)
874 * cause too much network io
875 *
876 * [89-fb] Using pread for non-compaction read request
877 * https://issues.apache.org/jira/browse/HBASE-7266
878 *
879 * On the other hand, if setting it true, we would do
880 * openScanner,next,closeScanner in one RPC call. It means the better
881 * performance for small scan. [HBASE-9488].
882 *
883 * Generally, if the scan range is within one data block(64KB), it could be
884 * considered as a small scan.
885 *
886 * @param small
887 */
888 public Scan setSmall(boolean small) {
889 this.small = small;
890 return this;
891 }
892
893 /**
894 * Get whether this scan is a small scan
895 * @return true if small scan
896 */
897 public boolean isSmall() {
898 return small;
899 }
900
901 @Override
902 public Scan setAttribute(String name, byte[] value) {
903 return (Scan) super.setAttribute(name, value);
904 }
905
906 @Override
907 public Scan setId(String id) {
908 return (Scan) super.setId(id);
909 }
910
911 @Override
912 public Scan setAuthorizations(Authorizations authorizations) {
913 return (Scan) super.setAuthorizations(authorizations);
914 }
915
916 @Override
917 public Scan setACL(Map<String, Permission> perms) {
918 return (Scan) super.setACL(perms);
919 }
920
921 @Override
922 public Scan setACL(String user, Permission perms) {
923 return (Scan) super.setACL(user, perms);
924 }
925
926 @Override
927 public Scan setConsistency(Consistency consistency) {
928 return (Scan) super.setConsistency(consistency);
929 }
930
931 @Override
932 public Scan setReplicaId(int Id) {
933 return (Scan) super.setReplicaId(Id);
934 }
935
936 @Override
937 public Scan setIsolationLevel(IsolationLevel level) {
938 return (Scan) super.setIsolationLevel(level);
939 }
940
941 /**
942 * Enable collection of {@link ScanMetrics}. For advanced users.
943 * @param enabled Set to true to enable accumulating scan metrics
944 */
945 public Scan setScanMetricsEnabled(final boolean enabled) {
946 setAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE, Bytes.toBytes(Boolean.valueOf(enabled)));
947 return this;
948 }
949
950 /**
951 * @return True if collection of scan metrics is enabled. For advanced users.
952 */
953 public boolean isScanMetricsEnabled() {
954 byte[] attr = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE);
955 return attr == null ? false : Bytes.toBoolean(attr);
956 }
957
958 /**
959 * @return Metrics on this Scan, if metrics were enabled.
960 * @see #setScanMetricsEnabled(boolean)
961 */
962 public ScanMetrics getScanMetrics() {
963 byte [] bytes = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_DATA);
964 if (bytes == null) return null;
965 return ProtobufUtil.toScanMetrics(bytes);
966 }
967
968 public Boolean isAsyncPrefetch() {
969 return asyncPrefetch;
970 }
971
972 public Scan setAsyncPrefetch(boolean asyncPrefetch) {
973 this.asyncPrefetch = asyncPrefetch;
974 return this;
975 }
976 }