Author: stack
Date: Tue Sep 27 18:42:31 2011
New Revision: 1176514
URL: http://svn.apache.org/viewvc?rev=1176514&view=rev
Log:
HBASE-4433: avoid extra next (potentially a seek) if done with column/row
Modified:
hbase/branches/0.92/CHANGES.txt
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ExplicitColumnTracker.java
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ScanQueryMatcher.java
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestBlocksRead.java
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestExplicitColumnTracker.java
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestQueryMatcher.java
Modified: hbase/branches/0.92/CHANGES.txt
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/CHANGES.txt?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
--- hbase/branches/0.92/CHANGES.txt (original)
+++ hbase/branches/0.92/CHANGES.txt Tue Sep 27 18:42:31 2011
@@ -634,6 +634,8 @@ Release 0.90.5 - Unreleased
(Lars Hofhansl)
HBASE-4295 rowcounter does not return the correct number of rows in
certain circumstances (David Revell)
+ HBASE-4433 avoid extra next (potentially a seek) if done with column/row
+ (Kannan Muthukkaruppan)
IMPROVEMENT
HBASE-4205 Enhance HTable javadoc (Eric Charles)
Modified:
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ExplicitColumnTracker.java
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ExplicitColumnTracker.java?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
---
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ExplicitColumnTracker.java
(original)
+++
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ExplicitColumnTracker.java
Tue Sep 27 18:42:31 2011
@@ -50,6 +50,12 @@ public class ExplicitColumnTracker imple
private final int maxVersions;
private final int minVersions;
+
+ /**
+ * Contains the list of columns that the ExplicitColumnTracker is tracking.
+ * Each ColumnCount instance also tracks how many versions of the requested
+ * column have been returned.
+ */
private final List<ColumnCount> columns;
private final List<ColumnCount> columnsToReuse;
private int index;
@@ -127,13 +133,22 @@ public class ExplicitColumnTracker imple
int count = this.column.increment();
if(count >= maxVersions || (count >= minVersions &&
isExpired(timestamp))) {
// Done with versions for this column
+ // Note: because we are done with this column, and are removing
+ // it from columns, we don't do a ++this.index. The index stays
+ // the same but the columns have shifted within the array such
+ // that index now points to the next column we are interested in.
this.columns.remove(this.index);
+
resetTS();
- if(this.columns.size() == this.index) {
- // Will not hit any more columns in this storefile
+ if (this.columns.size() == this.index) {
+ // We have served all the requested columns.
this.column = null;
+ return ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_ROW;
} else {
+ // We are done with current column; advance to next column
+ // of interest.
this.column = this.columns.get(this.index);
+ return ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL;
}
} else {
setTS(timestamp);
@@ -144,15 +159,18 @@ public class ExplicitColumnTracker imple
resetTS();
if (ret > 0) {
- // Specified column is smaller than the current, skip to next column.
+ // The current KV is smaller than the column the ExplicitColumnTracker
+ // is interested in, so seek to that column of interest.
return ScanQueryMatcher.MatchCode.SEEK_NEXT_COL;
}
- // Specified column is bigger than current column
- // Move down current column and check again
- if(ret <= -1) {
- if(++this.index >= this.columns.size()) {
- // No more to match, do not include, done with storefile
+ // The current KV is bigger than the column the ExplicitColumnTracker
+ // is interested in. That means there is no more data for the column
+ // of interest. Advance the ExplicitColumnTracker state to next
+ // column of interest, and check again.
+ if (ret <= -1) {
+ if (++this.index >= this.columns.size()) {
+ // No more to match, do not include, done with this row.
return ScanQueryMatcher.MatchCode.SEEK_NEXT_ROW; // done_row
}
// This is the recursive case.
Modified:
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ScanQueryMatcher.java
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ScanQueryMatcher.java?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
---
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ScanQueryMatcher.java
(original)
+++
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/ScanQueryMatcher.java
Tue Sep 27 18:42:31 2011
@@ -352,5 +352,15 @@ public class ScanQueryMatcher {
* Seek to next key which is given as hint.
*/
SEEK_NEXT_USING_HINT,
+
+ /**
+ * Include KeyValue and done with column, seek to next.
+ */
+ INCLUDE_AND_SEEK_NEXT_COL,
+
+ /**
+ * Include KeyValue and done with row, seek to next.
+ */
+ INCLUDE_AND_SEEK_NEXT_ROW,
}
}
Modified:
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
---
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java
(original)
+++
hbase/branches/0.92/src/main/java/org/apache/hadoop/hbase/regionserver/StoreScanner.java
Tue Sep 27 18:42:31 2011
@@ -258,8 +258,23 @@ class StoreScanner implements KeyValueSc
//DebugPrint.println("SS peek kv = " + kv + " with qcode = " + qcode);
switch(qcode) {
case INCLUDE:
+ case INCLUDE_AND_SEEK_NEXT_ROW:
+ case INCLUDE_AND_SEEK_NEXT_COL:
+
results.add(copyKv);
- this.heap.next();
+
+ if (qcode == ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_ROW) {
+ if (!matcher.moreRowsMayExistAfter(kv)) {
+ outResult.addAll(results);
+ return false;
+ }
+ reseek(matcher.getKeyForNextRow(kv));
+ } else if (qcode ==
ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL) {
+ reseek(matcher.getKeyForNextColumn(kv));
+ } else {
+ this.heap.next();
+ }
+
if (limit > 0 && (results.size() == limit)) {
break LOOP;
}
Modified:
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestBlocksRead.java
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestBlocksRead.java?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
---
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestBlocksRead.java
(original)
+++
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestBlocksRead.java
Tue Sep 27 18:42:31 2011
@@ -201,8 +201,8 @@ public class TestBlocksRead extends HBas
putData(FAMILY, "row", "col7", 7);
region.flushcache();
- // Expected block reads: 2
- kvs = getData(FAMILY, "row", "col1", 2);
+ // Expected block reads: 1
+ kvs = getData(FAMILY, "row", "col1", 1);
assertEquals(1, kvs.length);
verifyData(kvs[0], "row", "col1", 1);
@@ -218,8 +218,8 @@ public class TestBlocksRead extends HBas
verifyData(kvs[0], "row", "col2", 2);
verifyData(kvs[1], "row", "col3", 3);
- // Expected block reads: 4
- kvs = getData(FAMILY, "row", Arrays.asList("col5"), 4);
+ // Expected block reads: 3
+ kvs = getData(FAMILY, "row", Arrays.asList("col5"), 3);
assertEquals(1, kvs.length);
verifyData(kvs[0], "row", "col5", 5);
}
@@ -248,15 +248,16 @@ public class TestBlocksRead extends HBas
putData(FAMILY, "row", "col2", 4);
region.flushcache();
- // Baseline expected blocks read: 3
- kvs = getData(FAMILY, "row", Arrays.asList("col1"), 3);
+ // Baseline expected blocks read: 2
+ kvs = getData(FAMILY, "row", Arrays.asList("col1"), 2);
assertEquals(1, kvs.length);
verifyData(kvs[0], "row", "col1", 3);
- // Baseline expected blocks read: 5
+ // Baseline expected blocks read: 6
// This increase is a minor glitch due to: HBASE-4466. Once that
- // is fixed this will drop back. The extra access will be a cache hit.
- kvs = getData(FAMILY, "row", Arrays.asList("col1", "col2"), 5);
+ // is fixed this will drop back. The extra access will be a cache
+ // hit.
+ kvs = getData(FAMILY, "row", Arrays.asList("col1", "col2"), 6);
assertEquals(2, kvs.length);
verifyData(kvs[0], "row", "col1", 3);
verifyData(kvs[1], "row", "col2", 4);
@@ -271,8 +272,8 @@ public class TestBlocksRead extends HBas
verifyData(kvs[0], "row", "col3", 5);
// Get a column from older file.
- // Baseline expected blocks read: 4
- kvs = getData(FAMILY, "row", Arrays.asList("col1"), 4);
+ // Baseline expected blocks read: 3
+ kvs = getData(FAMILY, "row", Arrays.asList("col1"), 3);
assertEquals(1, kvs.length);
verifyData(kvs[0], "row", "col1", 3);
@@ -290,10 +291,12 @@ public class TestBlocksRead extends HBas
kvs = getData(FAMILY, "row", Arrays.asList("col1", "col2", "col3"), 6);
assertEquals(0, kvs.length);
- // File 5: Delete with post data timestamp and insert some older
- // date in new files.
+ // File 5: Delete
deleteFamily(FAMILY, "row", 10);
region.flushcache();
+
+ // File 6: some more puts, but with timestamps older than the
+ // previous delete.
putData(FAMILY, "row", "col1", 7);
putData(FAMILY, "row", "col2", 8);
putData(FAMILY, "row", "col3", 9);
@@ -303,14 +306,17 @@ public class TestBlocksRead extends HBas
kvs = getData(FAMILY, "row", Arrays.asList("col1", "col2", "col3"), 10);
assertEquals(0, kvs.length);
- // File 6: Put back new data
+ // File 7: Put back new data
putData(FAMILY, "row", "col1", 11);
putData(FAMILY, "row", "col2", 12);
putData(FAMILY, "row", "col3", 13);
region.flushcache();
- // Baseline expected blocks read: 13
- kvs = getData(FAMILY, "row", Arrays.asList("col1", "col2", "col3"), 13);
+ // Baseline expected blocks read: 21
+ // This increase is a minor glitch due to: HBASE-4466. Once that
+ // is fixed this will drop back. The extra access will be a cache
+ // hit. The test case only has 13 blocks altogther!
+ kvs = getData(FAMILY, "row", Arrays.asList("col1", "col2", "col3"), 21);
assertEquals(3, kvs.length);
verifyData(kvs[0], "row", "col1", 11);
verifyData(kvs[1], "row", "col2", 12);
Modified:
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestExplicitColumnTracker.java
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestExplicitColumnTracker.java?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
---
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestExplicitColumnTracker.java
(original)
+++
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestExplicitColumnTracker.java
Tue Sep 27 18:42:31 2011
@@ -77,11 +77,11 @@ public class TestExplicitColumnTracker e
columns.add(col2);
columns.add(col4);
List<MatchCode> expected = new ArrayList<ScanQueryMatcher.MatchCode>();
- expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
- expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
- expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_ROW);
+ expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL); // col1
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL); // col2
+ expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL); //
col3
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_ROW); // col4
+ expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_ROW); //
col5
int maxVersions = 1;
//Create "Scanner"
@@ -111,16 +111,16 @@ public class TestExplicitColumnTracker e
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE); //
col2; 1st version
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL); //
col2; 2nd version
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE); //
col4; 1st version
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_ROW); //
col4; 2nd version
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_ROW);
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_ROW);
Modified:
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestQueryMatcher.java
URL:
http://svn.apache.org/viewvc/hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestQueryMatcher.java?rev=1176514&r1=1176513&r2=1176514&view=diff
==============================================================================
---
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestQueryMatcher.java
(original)
+++
hbase/branches/0.92/src/test/java/org/apache/hadoop/hbase/regionserver/TestQueryMatcher.java
Tue Sep 27 18:42:31 2011
@@ -88,10 +88,10 @@ public class TestQueryMatcher extends HB
//Expected result
List<MatchCode> expected = new ArrayList<ScanQueryMatcher.MatchCode>();
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL);
expected.add(ScanQueryMatcher.MatchCode.SEEK_NEXT_COL);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
- expected.add(ScanQueryMatcher.MatchCode.INCLUDE);
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL);
+ expected.add(ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_ROW);
expected.add(ScanQueryMatcher.MatchCode.DONE);
// 2,4,5
@@ -182,9 +182,9 @@ public class TestQueryMatcher extends HB
long testTTL = 1000;
MatchCode [] expected = new MatchCode[] {
ScanQueryMatcher.MatchCode.SEEK_NEXT_COL,
- ScanQueryMatcher.MatchCode.INCLUDE,
+ ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL,
ScanQueryMatcher.MatchCode.SEEK_NEXT_COL,
- ScanQueryMatcher.MatchCode.INCLUDE,
+ ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL,
ScanQueryMatcher.MatchCode.SEEK_NEXT_ROW,
ScanQueryMatcher.MatchCode.DONE
};
