Jackie-Jiang commented on a change in pull request #7544:
URL: https://github.com/apache/pinot/pull/7544#discussion_r725178932
##########
File path:
pinot-core/src/main/java/org/apache/pinot/core/operator/InstanceResponseOperator.java
##########
@@ -19,61 +19,64 @@
package org.apache.pinot.core.operator;
import java.util.List;
-import org.apache.pinot.common.utils.DataTable;
import org.apache.pinot.common.utils.DataTable.MetadataKey;
-import org.apache.pinot.core.common.Operator;
import org.apache.pinot.core.operator.blocks.InstanceResponseBlock;
import org.apache.pinot.core.operator.blocks.IntermediateResultsBlock;
+import org.apache.pinot.core.operator.combine.BaseCombineOperator;
+import org.apache.pinot.core.query.request.context.ThreadTimer;
import org.apache.pinot.segment.spi.FetchContext;
import org.apache.pinot.segment.spi.IndexSegment;
public class InstanceResponseOperator extends
BaseOperator<InstanceResponseBlock> {
private static final String OPERATOR_NAME = "InstanceResponseOperator";
- private final Operator _operator;
+ private final BaseCombineOperator _combineOperator;
private final List<IndexSegment> _indexSegments;
private final List<FetchContext> _fetchContexts;
private final int _fetchContextSize;
- public InstanceResponseOperator(Operator combinedOperator,
List<IndexSegment> indexSegments,
+ public InstanceResponseOperator(BaseCombineOperator combinedOperator,
List<IndexSegment> indexSegments,
List<FetchContext> fetchContexts) {
- _operator = combinedOperator;
+ _combineOperator = combinedOperator;
_indexSegments = indexSegments;
_fetchContexts = fetchContexts;
_fetchContextSize = fetchContexts.size();
}
@Override
protected InstanceResponseBlock getNextBlock() {
- long startWallClockTimeNs = System.nanoTime();
+ if (ThreadTimer.isThreadCpuTimeMeasurementEnabled()) {
+ long startWallClockTimeNs = System.nanoTime();
+ IntermediateResultsBlock intermediateResultsBlock = getCombinedResults();
+ InstanceResponseBlock instanceResponseBlock = new
InstanceResponseBlock(intermediateResultsBlock);
+ long totalWallClockTimeNs = System.nanoTime() - startWallClockTimeNs;
- IntermediateResultsBlock intermediateResultsBlock;
+ /*
+ * If/when the threadCpuTime based instrumentation is done for other
parts of execution (planning, pruning etc),
+ * we will have to change the wallClockTime computation accordingly.
Right now everything under
+ * InstanceResponseOperator is the one that is instrumented with
threadCpuTime.
+ */
Review comment:
Not introduced in this PR, but based on my understanding, this comment
wants to indicate that currently we only count the CPU time for the physical
plan execution (`CombineOperator.nextBlock()`). Once other query execution
steps are included, we should also count them when calculating the wall clock
time. @mqliang Is this correct?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
