blob: e23a1232c7b3db113d85e020ae7c20fa1da86cfb [file] [log] [blame]
/*
* 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.iotdb.db.mpp.execution.operator.process;
import org.apache.iotdb.db.mpp.aggregation.Aggregator;
import org.apache.iotdb.db.mpp.aggregation.timerangeiterator.ITimeRangeIterator;
import org.apache.iotdb.db.mpp.execution.operator.Operator;
import org.apache.iotdb.db.mpp.execution.operator.OperatorContext;
import org.apache.iotdb.db.mpp.execution.operator.window.IWindow;
import org.apache.iotdb.db.mpp.execution.operator.window.IWindowManager;
import org.apache.iotdb.db.mpp.execution.operator.window.TimeWindowManager;
import java.util.List;
import static org.apache.iotdb.db.mpp.execution.operator.AggregationUtil.appendAggregationResult;
import static org.apache.iotdb.db.mpp.execution.operator.AggregationUtil.isAllAggregatorsHasFinalResult;
/**
* RawDataAggregationOperator is used to process raw data tsBlock input calculating using value
* filter. It's possible that there is more than one tsBlock input in one time interval. And it's
* also possible that one tsBlock can cover multiple time intervals too.
*
* <p>Since raw data query with value filter is processed by FilterOperator above TimeJoinOperator,
* there we can see RawDataAggregateOperator as a one-to-one(one input, ont output) operator.
*
* <p>Return aggregation result in many time intervals once.
*/
public class RawDataAggregationOperator extends SingleInputAggregationOperator {
private final IWindowManager windowManager;
public RawDataAggregationOperator(
OperatorContext operatorContext,
List<Aggregator> aggregators,
ITimeRangeIterator timeRangeIterator,
Operator child,
boolean ascending,
long maxReturnSize) {
super(operatorContext, aggregators, child, ascending, timeRangeIterator, maxReturnSize);
this.windowManager = new TimeWindowManager(timeRangeIterator);
}
private boolean hasMoreData() {
return inputTsBlock != null || child.hasNext();
}
@Override
public boolean hasNext() {
return windowManager.hasNext(hasMoreData());
}
@Override
protected boolean calculateNextAggregationResult() {
while (!calculateFromRawData()) {
inputTsBlock = null;
// NOTE: child.next() can only be invoked once
if (child.hasNext() && canCallNext) {
inputTsBlock = child.next();
canCallNext = false;
} else if (child.hasNext()) {
// if child still has next but can't be invoked now
return false;
} else {
// If there are no points belong to last window, the last window will not
// initialize window and aggregators
if (!windowManager.isCurWindowInit()) {
initWindowAndAggregators();
}
break;
}
}
updateResultTsBlock();
// Step into next window
windowManager.next();
return true;
}
private boolean calculateFromRawData() {
// if window is not initialized, we should init window status and reset aggregators
if (!windowManager.isCurWindowInit() && !skipPreviousWindowAndInitCurWindow()) {
return false;
}
// If current window has been initialized, we should judge whether inputTsBlock is empty
if (inputTsBlock == null || inputTsBlock.isEmpty()) {
return false;
}
if (windowManager.satisfiedCurWindow(inputTsBlock)) {
int lastReadRowIndex = 0;
for (Aggregator aggregator : aggregators) {
// Current agg method has been calculated
if (aggregator.hasFinalResult()) {
continue;
}
lastReadRowIndex = Math.max(lastReadRowIndex, aggregator.processTsBlock(inputTsBlock));
}
if (lastReadRowIndex >= inputTsBlock.getPositionCount()) {
inputTsBlock = null;
// For the last index of TsBlock, if we can know the aggregation calculation is over
// we can directly updateResultTsBlock and return true
return isAllAggregatorsHasFinalResult(aggregators);
} else {
inputTsBlock = inputTsBlock.subTsBlock(lastReadRowIndex);
return true;
}
}
boolean isTsBlockOutOfBound = windowManager.isTsBlockOutOfBound(inputTsBlock);
return isAllAggregatorsHasFinalResult(aggregators) || isTsBlockOutOfBound;
}
@Override
protected void updateResultTsBlock() {
appendAggregationResult(resultTsBlockBuilder, aggregators, windowManager.currentOutputTime());
}
private boolean skipPreviousWindowAndInitCurWindow() {
// Before we initialize windowManager and aggregators, we should ensure that we have consumed
// all points belong to previous window
inputTsBlock = windowManager.skipPointsOutOfCurWindow(inputTsBlock);
// After skipping, if tsBlock is empty, we cannot ensure that we have consumed all points belong
// to previous window, we should go back to calculateNextAggregationResult() to get a new
// tsBlock
if (inputTsBlock == null || inputTsBlock.isEmpty()) {
return false;
}
// If we have consumed all points belong to previous window, we can initialize current window
// and aggregators
initWindowAndAggregators();
return true;
}
private void initWindowAndAggregators() {
windowManager.initCurWindow(inputTsBlock);
IWindow curWindow = windowManager.getCurWindow();
for (Aggregator aggregator : aggregators) {
aggregator.updateWindow(curWindow);
}
}
}