View Javadoc

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     TimeRange ctr = scan.getTimeRange();
234     tr = new TimeRange(ctr.getMin(), ctr.getMax());
235     Map<byte[], NavigableSet<byte[]>> fams = scan.getFamilyMap();
236     for (Map.Entry<byte[],NavigableSet<byte[]>> entry : fams.entrySet()) {
237       byte [] fam = entry.getKey();
238       NavigableSet<byte[]> cols = entry.getValue();
239       if (cols != null && cols.size() > 0) {
240         for (byte[] col : cols) {
241           addColumn(fam, col);
242         }
243       } else {
244         addFamily(fam);
245       }
246     }
247     for (Map.Entry<String, byte[]> attr : scan.getAttributesMap().entrySet()) {
248       setAttribute(attr.getKey(), attr.getValue());
249     }
250   }
251 
252   /**
253    * Builds a scan object with the same specs as get.
254    * @param get get to model scan after
255    */
256   public Scan(Get get) {
257     this.startRow = get.getRow();
258     this.stopRow = get.getRow();
259     this.filter = get.getFilter();
260     this.cacheBlocks = get.getCacheBlocks();
261     this.maxVersions = get.getMaxVersions();
262     this.storeLimit = get.getMaxResultsPerColumnFamily();
263     this.storeOffset = get.getRowOffsetPerColumnFamily();
264     this.tr = get.getTimeRange();
265     this.familyMap = get.getFamilyMap();
266     this.getScan = true;
267     this.asyncPrefetch = false;
268     this.consistency = get.getConsistency();
269     for (Map.Entry<String, byte[]> attr : get.getAttributesMap().entrySet()) {
270       setAttribute(attr.getKey(), attr.getValue());
271     }
272   }
273 
274   public boolean isGetScan() {
275     return this.getScan || isStartRowAndEqualsStopRow();
276   }
277 
278   private boolean isStartRowAndEqualsStopRow() {
279     return this.startRow != null && this.startRow.length > 0 &&
280         Bytes.equals(this.startRow, this.stopRow);
281   }
282   /**
283    * Get all columns from the specified family.
284    * <p>
285    * Overrides previous calls to addColumn for this family.
286    * @param family family name
287    * @return this
288    */
289   public Scan addFamily(byte [] family) {
290     familyMap.remove(family);
291     familyMap.put(family, null);
292     return this;
293   }
294 
295   /**
296    * Get the column from the specified family with the specified qualifier.
297    * <p>
298    * Overrides previous calls to addFamily for this family.
299    * @param family family name
300    * @param qualifier column qualifier
301    * @return this
302    */
303   public Scan addColumn(byte [] family, byte [] qualifier) {
304     NavigableSet<byte []> set = familyMap.get(family);
305     if(set == null) {
306       set = new TreeSet<byte []>(Bytes.BYTES_COMPARATOR);
307     }
308     if (qualifier == null) {
309       qualifier = HConstants.EMPTY_BYTE_ARRAY;
310     }
311     set.add(qualifier);
312     familyMap.put(family, set);
313     return this;
314   }
315 
316   /**
317    * Get versions of columns only within the specified timestamp range,
318    * [minStamp, maxStamp).  Note, default maximum versions to return is 1.  If
319    * your time range spans more than one version and you want all versions
320    * returned, up the number of versions beyond the default.
321    * @param minStamp minimum timestamp value, inclusive
322    * @param maxStamp maximum timestamp value, exclusive
323    * @throws IOException if invalid time range
324    * @see #setMaxVersions()
325    * @see #setMaxVersions(int)
326    * @return this
327    */
328   public Scan setTimeRange(long minStamp, long maxStamp)
329   throws IOException {
330     tr = new TimeRange(minStamp, maxStamp);
331     return this;
332   }
333 
334   /**
335    * Get versions of columns with the specified timestamp. Note, default maximum
336    * versions to return is 1.  If your time range spans more than one version
337    * and you want all versions returned, up the number of versions beyond the
338    * defaut.
339    * @param timestamp version timestamp
340    * @see #setMaxVersions()
341    * @see #setMaxVersions(int)
342    * @return this
343    */
344   public Scan setTimeStamp(long timestamp)
345   throws IOException {
346     try {
347       tr = new TimeRange(timestamp, timestamp+1);
348     } catch(IOException e) {
349       // This should never happen, unless integer overflow or something extremely wrong...
350       LOG.error("TimeRange failed, likely caused by integer overflow. ", e);
351       throw e;
352     }
353     return this;
354   }
355 
356   /**
357    * Set the start row of the scan.
358    * <p>
359    * If the specified row does not exist, the Scanner will start from the
360    * next closest row after the specified row.
361    * @param startRow row to start scanner at or after
362    * @return this
363    */
364   public Scan setStartRow(byte [] startRow) {
365     this.startRow = startRow;
366     return this;
367   }
368 
369   /**
370    * Set the stop row of the scan.
371    * @param stopRow row to end at (exclusive)
372    * <p>
373    * The scan will include rows that are lexicographically less than
374    * the provided stopRow.
375    * <p><b>Note:</b> When doing a filter for a rowKey <u>Prefix</u>
376    * use {@link #setRowPrefixFilter(byte[])}.
377    * The 'trailing 0' will not yield the desired result.</p>
378    * @return this
379    */
380   public Scan setStopRow(byte [] stopRow) {
381     this.stopRow = stopRow;
382     return this;
383   }
384 
385   /**
386    * <p>Set a filter (using stopRow and startRow) so the result set only contains rows where the
387    * rowKey starts with the specified prefix.</p>
388    * <p>This is a utility method that converts the desired rowPrefix into the appropriate values
389    * for the startRow and stopRow to achieve the desired result.</p>
390    * <p>This can safely be used in combination with setFilter.</p>
391    * <p><b>NOTE: Doing a {@link #setStartRow(byte[])} and/or {@link #setStopRow(byte[])}
392    * after this method will yield undefined results.</b></p>
393    * @param rowPrefix the prefix all rows must start with. (Set <i>null</i> to remove the filter.)
394    * @return this
395    */
396   public Scan setRowPrefixFilter(byte[] rowPrefix) {
397     if (rowPrefix == null) {
398       setStartRow(HConstants.EMPTY_START_ROW);
399       setStopRow(HConstants.EMPTY_END_ROW);
400     } else {
401       this.setStartRow(rowPrefix);
402       this.setStopRow(calculateTheClosestNextRowKeyForPrefix(rowPrefix));
403     }
404     return this;
405   }
406 
407   /**
408    * <p>When scanning for a prefix the scan should stop immediately after the the last row that
409    * has the specified prefix. This method calculates the closest next rowKey immediately following
410    * the given rowKeyPrefix.</p>
411    * <p><b>IMPORTANT: This converts a rowKey<u>Prefix</u> into a rowKey</b>.</p>
412    * <p>If the prefix is an 'ASCII' string put into a byte[] then this is easy because you can
413    * simply increment the last byte of the array.
414    * But if your application uses real binary rowids you may run into the scenario that your
415    * prefix is something like:</p>
416    * &nbsp;&nbsp;&nbsp;<b>{ 0x12, 0x23, 0xFF, 0xFF }</b><br/>
417    * Then this stopRow needs to be fed into the actual scan<br/>
418    * &nbsp;&nbsp;&nbsp;<b>{ 0x12, 0x24 }</b> (Notice that it is shorter now)<br/>
419    * This method calculates the correct stop row value for this usecase.
420    *
421    * @param rowKeyPrefix the rowKey<u>Prefix</u>.
422    * @return the closest next rowKey immediately following the given rowKeyPrefix.
423    */
424   private byte[] calculateTheClosestNextRowKeyForPrefix(byte[] rowKeyPrefix) {
425     // Essentially we are treating it like an 'unsigned very very long' and doing +1 manually.
426     // Search for the place where the trailing 0xFFs start
427     int offset = rowKeyPrefix.length;
428     while (offset > 0) {
429       if (rowKeyPrefix[offset - 1] != (byte) 0xFF) {
430         break;
431       }
432       offset--;
433     }
434 
435     if (offset == 0) {
436       // We got an 0xFFFF... (only FFs) stopRow value which is
437       // the last possible prefix before the end of the table.
438       // So set it to stop at the 'end of the table'
439       return HConstants.EMPTY_END_ROW;
440     }
441 
442     // Copy the right length of the original
443     byte[] newStopRow = Arrays.copyOfRange(rowKeyPrefix, 0, offset);
444     // And increment the last one
445     newStopRow[newStopRow.length - 1]++;
446     return newStopRow;
447   }
448 
449   /**
450    * Get all available versions.
451    * @return this
452    */
453   public Scan setMaxVersions() {
454     this.maxVersions = Integer.MAX_VALUE;
455     return this;
456   }
457 
458   /**
459    * Get up to the specified number of versions of each column.
460    * @param maxVersions maximum versions for each column
461    * @return this
462    */
463   public Scan setMaxVersions(int maxVersions) {
464     this.maxVersions = maxVersions;
465     return this;
466   }
467 
468   /**
469    * Set the maximum number of values to return for each call to next()
470    * @param batch the maximum number of values
471    */
472   public Scan setBatch(int batch) {
473     if (this.hasFilter() && this.filter.hasFilterRow()) {
474       throw new IncompatibleFilterException(
475         "Cannot set batch on a scan using a filter" +
476         " that returns true for filter.hasFilterRow");
477     }
478     this.batch = batch;
479     return this;
480   }
481 
482   /**
483    * Set the maximum number of values to return per row per Column Family
484    * @param limit the maximum number of values returned / row / CF
485    */
486   public Scan setMaxResultsPerColumnFamily(int limit) {
487     this.storeLimit = limit;
488     return this;
489   }
490 
491   /**
492    * Set offset for the row per Column Family.
493    * @param offset is the number of kvs that will be skipped.
494    */
495   public Scan setRowOffsetPerColumnFamily(int offset) {
496     this.storeOffset = offset;
497     return this;
498   }
499 
500   /**
501    * Set the number of rows for caching that will be passed to scanners.
502    * If not set, the Configuration setting {@link HConstants#HBASE_CLIENT_SCANNER_CACHING} will
503    * apply.
504    * Higher caching values will enable faster scanners but will use more memory.
505    * @param caching the number of rows for caching
506    */
507   public Scan setCaching(int caching) {
508     this.caching = caching;
509     return this;
510   }
511 
512   /**
513    * @return the maximum result size in bytes. See {@link #setMaxResultSize(long)}
514    */
515   public long getMaxResultSize() {
516     return maxResultSize;
517   }
518 
519   /**
520    * Set the maximum result size. The default is -1; this means that no specific
521    * maximum result size will be set for this scan, and the global configured
522    * value will be used instead. (Defaults to unlimited).
523    *
524    * @param maxResultSize The maximum result size in bytes.
525    */
526   public Scan setMaxResultSize(long maxResultSize) {
527     this.maxResultSize = maxResultSize;
528     return this;
529   }
530 
531   @Override
532   public Scan setFilter(Filter filter) {
533     super.setFilter(filter);
534     return this;
535   }
536 
537   /**
538    * Setting the familyMap
539    * @param familyMap map of family to qualifier
540    * @return this
541    */
542   public Scan setFamilyMap(Map<byte [], NavigableSet<byte []>> familyMap) {
543     this.familyMap = familyMap;
544     return this;
545   }
546 
547   /**
548    * Getting the familyMap
549    * @return familyMap
550    */
551   public Map<byte [], NavigableSet<byte []>> getFamilyMap() {
552     return this.familyMap;
553   }
554 
555   /**
556    * @return the number of families in familyMap
557    */
558   public int numFamilies() {
559     if(hasFamilies()) {
560       return this.familyMap.size();
561     }
562     return 0;
563   }
564 
565   /**
566    * @return true if familyMap is non empty, false otherwise
567    */
568   public boolean hasFamilies() {
569     return !this.familyMap.isEmpty();
570   }
571 
572   /**
573    * @return the keys of the familyMap
574    */
575   public byte[][] getFamilies() {
576     if(hasFamilies()) {
577       return this.familyMap.keySet().toArray(new byte[0][0]);
578     }
579     return null;
580   }
581 
582   /**
583    * @return the startrow
584    */
585   public byte [] getStartRow() {
586     return this.startRow;
587   }
588 
589   /**
590    * @return the stoprow
591    */
592   public byte [] getStopRow() {
593     return this.stopRow;
594   }
595 
596   /**
597    * @return the max number of versions to fetch
598    */
599   public int getMaxVersions() {
600     return this.maxVersions;
601   }
602 
603   /**
604    * @return maximum number of values to return for a single call to next()
605    */
606   public int getBatch() {
607     return this.batch;
608   }
609 
610   /**
611    * @return maximum number of values to return per row per CF
612    */
613   public int getMaxResultsPerColumnFamily() {
614     return this.storeLimit;
615   }
616 
617   /**
618    * Method for retrieving the scan's offset per row per column
619    * family (#kvs to be skipped)
620    * @return row offset
621    */
622   public int getRowOffsetPerColumnFamily() {
623     return this.storeOffset;
624   }
625 
626   /**
627    * @return caching the number of rows fetched when calling next on a scanner
628    */
629   public int getCaching() {
630     return this.caching;
631   }
632 
633   /**
634    * @return TimeRange
635    */
636   public TimeRange getTimeRange() {
637     return this.tr;
638   }
639 
640   /**
641    * @return RowFilter
642    */
643   @Override
644   public Filter getFilter() {
645     return filter;
646   }
647 
648   /**
649    * @return true is a filter has been specified, false if not
650    */
651   public boolean hasFilter() {
652     return filter != null;
653   }
654 
655   /**
656    * Set whether blocks should be cached for this Scan.
657    * <p>
658    * This is true by default.  When true, default settings of the table and
659    * family are used (this will never override caching blocks if the block
660    * cache is disabled for that family or entirely).
661    *
662    * @param cacheBlocks if false, default settings are overridden and blocks
663    * will not be cached
664    */
665   public Scan setCacheBlocks(boolean cacheBlocks) {
666     this.cacheBlocks = cacheBlocks;
667     return this;
668   }
669 
670   /**
671    * Get whether blocks should be cached for this Scan.
672    * @return true if default caching should be used, false if blocks should not
673    * be cached
674    */
675   public boolean getCacheBlocks() {
676     return cacheBlocks;
677   }
678 
679   /**
680    * Set whether this scan is a reversed one
681    * <p>
682    * This is false by default which means forward(normal) scan.
683    *
684    * @param reversed if true, scan will be backward order
685    * @return this
686    */
687   public Scan setReversed(boolean reversed) {
688     this.reversed = reversed;
689     return this;
690   }
691 
692   /**
693    * Get whether this scan is a reversed one.
694    * @return true if backward scan, false if forward(default) scan
695    */
696   public boolean isReversed() {
697     return reversed;
698   }
699 
700   /**
701    * Setting whether the caller wants to see the partial results that may be returned from the
702    * server. By default this value is false and the complete results will be assembled client side
703    * before being delivered to the caller.
704    * @param allowPartialResults
705    * @return this
706    */
707   public Scan setAllowPartialResults(final boolean allowPartialResults) {
708     this.allowPartialResults = allowPartialResults;
709     return this;
710   }
711 
712   /**
713    * @return true when the constructor of this scan understands that the results they will see may
714    *         only represent a partial portion of a row. The entire row would be retrieved by
715    *         subsequent calls to {@link ResultScanner#next()}
716    */
717   public boolean getAllowPartialResults() {
718     return allowPartialResults;
719   }
720 
721   /**
722    * Set the value indicating whether loading CFs on demand should be allowed (cluster
723    * default is false). On-demand CF loading doesn't load column families until necessary, e.g.
724    * if you filter on one column, the other column family data will be loaded only for the rows
725    * that are included in result, not all rows like in normal case.
726    * With column-specific filters, like SingleColumnValueFilter w/filterIfMissing == true,
727    * this can deliver huge perf gains when there's a cf with lots of data; however, it can
728    * also lead to some inconsistent results, as follows:
729    * - if someone does a concurrent update to both column families in question you may get a row
730    *   that never existed, e.g. for { rowKey = 5, { cat_videos =&gt; 1 }, { video =&gt; "my cat" } }
731    *   someone puts rowKey 5 with { cat_videos =&gt; 0 }, { video =&gt; "my dog" }, concurrent scan
732    *   filtering on "cat_videos == 1" can get { rowKey = 5, { cat_videos =&gt; 1 },
733    *   { video =&gt; "my dog" } }.
734    * - if there's a concurrent split and you have more than 2 column families, some rows may be
735    *   missing some column families.
736    */
737   public Scan setLoadColumnFamiliesOnDemand(boolean value) {
738     this.loadColumnFamiliesOnDemand = value;
739     return this;
740   }
741 
742   /**
743    * Get the raw loadColumnFamiliesOnDemand setting; if it's not set, can be null.
744    */
745   public Boolean getLoadColumnFamiliesOnDemandValue() {
746     return this.loadColumnFamiliesOnDemand;
747   }
748 
749   /**
750    * Get the logical value indicating whether on-demand CF loading should be allowed.
751    */
752   public boolean doLoadColumnFamiliesOnDemand() {
753     return (this.loadColumnFamiliesOnDemand != null)
754       && this.loadColumnFamiliesOnDemand.booleanValue();
755   }
756 
757   /**
758    * Compile the table and column family (i.e. schema) information
759    * into a String. Useful for parsing and aggregation by debugging,
760    * logging, and administration tools.
761    * @return Map
762    */
763   @Override
764   public Map<String, Object> getFingerprint() {
765     Map<String, Object> map = new HashMap<String, Object>();
766     List<String> families = new ArrayList<String>();
767     if(this.familyMap.size() == 0) {
768       map.put("families", "ALL");
769       return map;
770     } else {
771       map.put("families", families);
772     }
773     for (Map.Entry<byte [], NavigableSet<byte[]>> entry :
774         this.familyMap.entrySet()) {
775       families.add(Bytes.toStringBinary(entry.getKey()));
776     }
777     return map;
778   }
779 
780   /**
781    * Compile the details beyond the scope of getFingerprint (row, columns,
782    * timestamps, etc.) into a Map along with the fingerprinted information.
783    * Useful for debugging, logging, and administration tools.
784    * @param maxCols a limit on the number of columns output prior to truncation
785    * @return Map
786    */
787   @Override
788   public Map<String, Object> toMap(int maxCols) {
789     // start with the fingerpring map and build on top of it
790     Map<String, Object> map = getFingerprint();
791     // map from families to column list replaces fingerprint's list of families
792     Map<String, List<String>> familyColumns =
793       new HashMap<String, List<String>>();
794     map.put("families", familyColumns);
795     // add scalar information first
796     map.put("startRow", Bytes.toStringBinary(this.startRow));
797     map.put("stopRow", Bytes.toStringBinary(this.stopRow));
798     map.put("maxVersions", this.maxVersions);
799     map.put("batch", this.batch);
800     map.put("caching", this.caching);
801     map.put("maxResultSize", this.maxResultSize);
802     map.put("cacheBlocks", this.cacheBlocks);
803     map.put("loadColumnFamiliesOnDemand", this.loadColumnFamiliesOnDemand);
804     List<Long> timeRange = new ArrayList<Long>();
805     timeRange.add(this.tr.getMin());
806     timeRange.add(this.tr.getMax());
807     map.put("timeRange", timeRange);
808     int colCount = 0;
809     // iterate through affected families and list out up to maxCols columns
810     for (Map.Entry<byte [], NavigableSet<byte[]>> entry :
811       this.familyMap.entrySet()) {
812       List<String> columns = new ArrayList<String>();
813       familyColumns.put(Bytes.toStringBinary(entry.getKey()), columns);
814       if(entry.getValue() == null) {
815         colCount++;
816         --maxCols;
817         columns.add("ALL");
818       } else {
819         colCount += entry.getValue().size();
820         if (maxCols <= 0) {
821           continue;
822         }
823         for (byte [] column : entry.getValue()) {
824           if (--maxCols <= 0) {
825             continue;
826           }
827           columns.add(Bytes.toStringBinary(column));
828         }
829       }
830     }
831     map.put("totalColumns", colCount);
832     if (this.filter != null) {
833       map.put("filter", this.filter.toString());
834     }
835     // add the id if set
836     if (getId() != null) {
837       map.put("id", getId());
838     }
839     return map;
840   }
841 
842   /**
843    * Enable/disable "raw" mode for this scan.
844    * If "raw" is enabled the scan will return all
845    * delete marker and deleted rows that have not
846    * been collected, yet.
847    * This is mostly useful for Scan on column families
848    * that have KEEP_DELETED_ROWS enabled.
849    * It is an error to specify any column when "raw" is set.
850    * @param raw True/False to enable/disable "raw" mode.
851    */
852   public Scan setRaw(boolean raw) {
853     setAttribute(RAW_ATTR, Bytes.toBytes(raw));
854     return this;
855   }
856 
857   /**
858    * @return True if this Scan is in "raw" mode.
859    */
860   public boolean isRaw() {
861     byte[] attr = getAttribute(RAW_ATTR);
862     return attr == null ? false : Bytes.toBoolean(attr);
863   }
864 
865 
866 
867   /**
868    * Set whether this scan is a small scan
869    * <p>
870    * Small scan should use pread and big scan can use seek + read
871    *
872    * seek + read is fast but can cause two problem (1) resource contention (2)
873    * cause too much network io
874    *
875    * [89-fb] Using pread for non-compaction read request
876    * https://issues.apache.org/jira/browse/HBASE-7266
877    *
878    * On the other hand, if setting it true, we would do
879    * openScanner,next,closeScanner in one RPC call. It means the better
880    * performance for small scan. [HBASE-9488].
881    *
882    * Generally, if the scan range is within one data block(64KB), it could be
883    * considered as a small scan.
884    *
885    * @param small
886    */
887   public Scan setSmall(boolean small) {
888     this.small = small;
889     return this;
890   }
891 
892   /**
893    * Get whether this scan is a small scan
894    * @return true if small scan
895    */
896   public boolean isSmall() {
897     return small;
898   }
899 
900   @Override
901   public Scan setAttribute(String name, byte[] value) {
902     return (Scan) super.setAttribute(name, value);
903   }
904 
905   @Override
906   public Scan setId(String id) {
907     return (Scan) super.setId(id);
908   }
909 
910   @Override
911   public Scan setAuthorizations(Authorizations authorizations) {
912     return (Scan) super.setAuthorizations(authorizations);
913   }
914 
915   @Override
916   public Scan setACL(Map<String, Permission> perms) {
917     return (Scan) super.setACL(perms);
918   }
919 
920   @Override
921   public Scan setACL(String user, Permission perms) {
922     return (Scan) super.setACL(user, perms);
923   }
924 
925   @Override
926   public Scan setConsistency(Consistency consistency) {
927     return (Scan) super.setConsistency(consistency);
928   }
929 
930   @Override
931   public Scan setReplicaId(int Id) {
932     return (Scan) super.setReplicaId(Id);
933   }
934 
935   @Override
936   public Scan setIsolationLevel(IsolationLevel level) {
937     return (Scan) super.setIsolationLevel(level);
938   }
939 
940   /**
941    * Utility that creates a Scan that will do a  small scan in reverse from passed row
942    * looking for next closest row.
943    * @param row
944    * @param family
945    * @return An instance of Scan primed with passed <code>row</code> and <code>family</code> to
946    * scan in reverse for one row only.
947    */
948   static Scan createGetClosestRowOrBeforeReverseScan(byte[] row) {
949     // Below does not work if you add in family; need to add the family qualifier that is highest
950     // possible family qualifier.  Do we have such a notion?  Would have to be magic.
951     Scan scan = new Scan(row);
952     scan.setSmall(true);
953     scan.setReversed(true);
954     scan.setCaching(1);
955     return scan;
956   }
957 
958   /**
959    * Enable collection of {@link ScanMetrics}. For advanced users.
960    * @param enabled Set to true to enable accumulating scan metrics
961    */
962   public Scan setScanMetricsEnabled(final boolean enabled) {
963     setAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE, Bytes.toBytes(Boolean.valueOf(enabled)));
964     return this;
965   }
966 
967   /**
968    * @return True if collection of scan metrics is enabled. For advanced users.
969    */
970   public boolean isScanMetricsEnabled() {
971     byte[] attr = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_ENABLE);
972     return attr == null ? false : Bytes.toBoolean(attr);
973   }
974 
975   /**
976    * @return Metrics on this Scan, if metrics were enabled.
977    * @see #setScanMetricsEnabled(boolean)
978    */
979   public ScanMetrics getScanMetrics() {
980     byte [] bytes = getAttribute(Scan.SCAN_ATTRIBUTES_METRICS_DATA);
981     if (bytes == null) return null;
982     return ProtobufUtil.toScanMetrics(bytes);
983   }
984 
985   public Boolean isAsyncPrefetch() {
986     return asyncPrefetch;
987   }
988 
989   public Scan setAsyncPrefetch(boolean asyncPrefetch) {
990     this.asyncPrefetch = asyncPrefetch;
991     return this;
992   }
993 }