[
https://issues.apache.org/jira/browse/DRILL-6381?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16647438#comment-16647438
]
ASF GitHub Bot commented on DRILL-6381:
---------------------------------------
vdiravka commented on a change in pull request #1466: DRILL-6381: Add support
for index based planning and execution
URL: https://github.com/apache/drill/pull/1466#discussion_r223664534
##########
File path:
exec/java-exec/src/main/java/org/apache/drill/exec/physical/impl/join/RowKeyJoinBatch.java
##########
@@ -0,0 +1,284 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+package org.apache.drill.exec.physical.impl.join;
+
+
+import java.util.List;
+
+import org.apache.commons.lang3.tuple.Pair;
+import org.apache.drill.exec.exception.OutOfMemoryException;
+import org.apache.drill.exec.exception.SchemaChangeException;
+import org.apache.drill.exec.ops.FragmentContext;
+import org.apache.drill.exec.physical.config.RowKeyJoinPOP;
+import org.apache.drill.exec.record.AbstractRecordBatch;
+import org.apache.drill.exec.record.RecordBatch;
+import org.apache.drill.exec.record.TransferPair;
+import org.apache.drill.exec.record.VectorWrapper;
+import org.apache.drill.exec.record.selection.SelectionVector2;
+import org.apache.drill.exec.record.selection.SelectionVector4;
+import org.apache.drill.exec.vector.SchemaChangeCallBack;
+import org.apache.drill.exec.vector.ValueVector;
+
+import org.apache.drill.shaded.guava.com.google.common.collect.Iterables;
+import org.apache.drill.shaded.guava.com.google.common.collect.Lists;
+
+
+public class RowKeyJoinBatch extends AbstractRecordBatch<RowKeyJoinPOP>
implements RowKeyJoin {
+ private static final org.slf4j.Logger logger =
org.slf4j.LoggerFactory.getLogger(RowKeyJoinBatch.class);
+
+ // primary table side record batch
+ private final RecordBatch left;
+
+ // index table side record batch
+ private final RecordBatch right;
+
+ private boolean hasRowKeyBatch;
+ private IterOutcome leftUpstream = IterOutcome.NONE;
+ private IterOutcome rightUpstream = IterOutcome.NONE;
+ private final List<TransferPair> transfers = Lists.newArrayList();
+ private int recordCount = 0;
+ private SchemaChangeCallBack callBack = new SchemaChangeCallBack();
+ private RowKeyJoinState rkJoinState = RowKeyJoinState.INITIAL;
+
+ public RowKeyJoinBatch(RowKeyJoinPOP config, FragmentContext context,
RecordBatch left, RecordBatch right)
+ throws OutOfMemoryException {
+ super(config, context, true /* need to build schema */);
+ this.left = left;
+ this.right = right;
+ this.hasRowKeyBatch = false;
+ }
+
+ @Override
+ public int getRecordCount() {
+ if (state == BatchState.DONE) {
+ return 0;
+ }
+ return recordCount;
+ }
+
+ @Override
+ public SelectionVector2 getSelectionVector2() {
+ throw new UnsupportedOperationException("RowKeyJoinBatch does not support
selection vector");
+ }
+
+ @Override
+ public SelectionVector4 getSelectionVector4() {
+ throw new UnsupportedOperationException("RowKeyJoinBatch does not support
selection vector");
+ }
+
+ @Override
+ protected void buildSchema() throws SchemaChangeException {
+ container.clear();
+
+ rightUpstream = next(right);
+
+ if (leftUpstream == IterOutcome.STOP || rightUpstream == IterOutcome.STOP)
{
+ state = BatchState.STOP;
+ return;
+ }
+
+ if (right.getRecordCount() > 0) {
+ // set the hasRowKeyBatch flag such that calling next() on the left input
+ // would see the correct status
+ hasRowKeyBatch = true;
+ }
+
+ leftUpstream = next(left);
+
+ if (leftUpstream == IterOutcome.OUT_OF_MEMORY || rightUpstream ==
IterOutcome.OUT_OF_MEMORY) {
+ state = BatchState.OUT_OF_MEMORY;
+ return;
+ }
+
+ for(final VectorWrapper<?> v : left) {
+ final TransferPair pair = v.getValueVector().makeTransferPair(
+ container.addOrGet(v.getField(), callBack));
+ transfers.add(pair);
+ }
+
+ container.buildSchema(left.getSchema().getSelectionVectorMode());
+ }
+
+ @Override
+ public IterOutcome innerNext() {
+ if (state == BatchState.DONE) {
+ return IterOutcome.NONE;
+ }
+ try {
+ if (state == BatchState.FIRST && left.getRecordCount() > 0) {
+ logger.debug("First batch, outputting the batch with {} records.",
left.getRecordCount());
+ // there is already a pending batch from left, output it
+ outputCurrentLeftBatch();
+ // Check if schema has changed (this is just to guard against
potential changes to the
+ // output schema by outputCurrentLeftBatch.
+ if (callBack.getSchemaChangedAndReset()) {
+ return IterOutcome.OK_NEW_SCHEMA;
+ }
+ return IterOutcome.OK;
+ }
+
+ rightUpstream = next(right);
+
+ logger.debug("right input IterOutcome: {}", rightUpstream);
+
+ switch(rightUpstream) {
+ case NONE:
+ case OUT_OF_MEMORY:
+ case STOP:
+ rkJoinState = RowKeyJoinState.DONE;
+ state = BatchState.DONE;
+ return rightUpstream;
+ case OK_NEW_SCHEMA:
+ case OK:
+ // we got a new batch from the right input, set this flag
+ // such that subsequent check by a scan would indicate availability
+ // of the row keys.
+ while((rightUpstream == IterOutcome.OK || rightUpstream ==
IterOutcome.OK_NEW_SCHEMA) &&
Review comment:
formatting
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> Add capability to do index based planning and execution
> -------------------------------------------------------
>
> Key: DRILL-6381
> URL: https://issues.apache.org/jira/browse/DRILL-6381
> Project: Apache Drill
> Issue Type: New Feature
> Components: Execution - Relational Operators, Query Planning &
> Optimization
> Reporter: Aman Sinha
> Assignee: Aman Sinha
> Priority: Major
> Fix For: 1.15.0
>
>
> If the underlying data source supports indexes (primary and secondary
> indexes), Drill should leverage those during planning and execution in order
> to improve query performance.
> On the planning side, Drill planner should be enhanced to provide an
> abstraction layer which express the index metadata and statistics. Further,
> a cost-based index selection is needed to decide which index(es) are
> suitable.
> On the execution side, appropriate operator enhancements would be needed to
> handle different categories of indexes such as covering, non-covering
> indexes, taking into consideration the index data may not be co-located with
> the primary table, i.e a global index.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
