/*
*
* 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.hadoop.hbase.coprocessor;
import java.io.IOException;
import java.nio.ByteBuffer;
import java.util.ArrayList;
import java.util.List;
import java.util.NavigableSet;
import org.apache.commons.logging.Log;
import org.apache.commons.logging.LogFactory;
import org.apache.hadoop.classification.InterfaceAudience;
import org.apache.hadoop.classification.InterfaceStability;
import org.apache.hadoop.hbase.Coprocessor;
import org.apache.hadoop.hbase.CoprocessorEnvironment;
import org.apache.hadoop.hbase.KeyValue;
import org.apache.hadoop.hbase.client.Scan;
import org.apache.hadoop.hbase.client.coprocessor.AggregationClient;
import org.apache.hadoop.hbase.filter.FirstKeyOnlyFilter;
import org.apache.hadoop.hbase.protobuf.ProtobufUtil;
import org.apache.hadoop.hbase.protobuf.ResponseConverter;
import org.apache.hadoop.hbase.protobuf.generated.AggregateProtos.AggregateArgument;
import org.apache.hadoop.hbase.protobuf.generated.AggregateProtos.AggregateResponse;
import org.apache.hadoop.hbase.protobuf.generated.AggregateProtos.AggregateService;
import org.apache.hadoop.hbase.regionserver.InternalScanner;
import com.google.protobuf.ByteString;
import com.google.protobuf.RpcCallback;
import com.google.protobuf.RpcController;
import com.google.protobuf.Service;
/**
* A concrete AggregateProtocol implementation. Its system level coprocessor
* that computes the aggregate function at a region level.
* @param <T>
* @param <S>
*/
@InterfaceAudience.Public
@InterfaceStability.Evolving
public class AggregateImplementation<T, S> extends AggregateService implements
CoprocessorService, Coprocessor {
protected static Log log = LogFactory.getLog(AggregateImplementation.class);
private RegionCoprocessorEnvironment env;
/**
* Gives the maximum for a given combination of column qualifier and column
* family, in the given row range as defined in the Scan object. In its
* current implementation, it takes one column family and one column qualifier
* (if provided). In case of null column qualifier, maximum value for the
* entire column family will be returned.
*/
@Override
public void getMax(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
InternalScanner scanner = null;
AggregateResponse response = null;
T max = null;
try {
ColumnInterpreter<T, S> ci = constructColumnInterpreterFromRequest(request);
T temp;
Scan scan = ProtobufUtil.toScan(request.getScan());
scanner = env.getRegion().getScanner(scan);
List<KeyValue> results = new ArrayList<KeyValue>();
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] qualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
qualifier = qualifiers.pollFirst();
}
// qualifier can be null.
boolean hasMoreRows = false;
do {
hasMoreRows = scanner.next(results);
for (KeyValue kv : results) {
temp = ci.getValue(colFamily, qualifier, kv);
max = (max == null || (temp != null && ci.compare(temp, max) > 0)) ? temp : max;
}
results.clear();
} while (hasMoreRows);
if (max != null) {
AggregateResponse.Builder builder = AggregateResponse.newBuilder();
builder.addFirstPart(ci.getProtoForCellType(max));
response = builder.build();
}
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
log.info("Maximum from this region is "
+ env.getRegion().getRegionNameAsString() + ": " + max);
done.run(response);
}
/**
* Gives the minimum for a given combination of column qualifier and column
* family, in the given row range as defined in the Scan object. In its
* current implementation, it takes one column family and one column qualifier
* (if provided). In case of null column qualifier, minimum value for the
* entire column family will be returned.
*/
@Override
public void getMin(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
AggregateResponse response = null;
InternalScanner scanner = null;
T min = null;
try {
ColumnInterpreter<T, S> ci = constructColumnInterpreterFromRequest(request);
T temp;
Scan scan = ProtobufUtil.toScan(request.getScan());
scanner = env.getRegion().getScanner(scan);
List<KeyValue> results = new ArrayList<KeyValue>();
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] qualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
qualifier = qualifiers.pollFirst();
}
boolean hasMoreRows = false;
do {
hasMoreRows = scanner.next(results);
for (KeyValue kv : results) {
temp = ci.getValue(colFamily, qualifier, kv);
min = (min == null || (temp != null && ci.compare(temp, min) < 0)) ? temp : min;
}
results.clear();
} while (hasMoreRows);
if (min != null) {
response = AggregateResponse.newBuilder().addFirstPart(
ci.getProtoForCellType(min)).build();
}
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
log.info("Minimum from this region is "
+ env.getRegion().getRegionNameAsString() + ": " + min);
done.run(response);
}
/**
* Gives the sum for a given combination of column qualifier and column
* family, in the given row range as defined in the Scan object. In its
* current implementation, it takes one column family and one column qualifier
* (if provided). In case of null column qualifier, sum for the entire column
* family will be returned.
*/
@Override
public void getSum(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
AggregateResponse response = null;
InternalScanner scanner = null;
long sum = 0l;
try {
ColumnInterpreter<T, S> ci = constructColumnInterpreterFromRequest(request);
S sumVal = null;
T temp;
Scan scan = ProtobufUtil.toScan(request.getScan());
scanner = env.getRegion().getScanner(scan);
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] qualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
qualifier = qualifiers.pollFirst();
}
List<KeyValue> results = new ArrayList<KeyValue>();
boolean hasMoreRows = false;
do {
hasMoreRows = scanner.next(results);
for (KeyValue kv : results) {
temp = ci.getValue(colFamily, qualifier, kv);
if (temp != null)
sumVal = ci.add(sumVal, ci.castToReturnType(temp));
}
results.clear();
} while (hasMoreRows);
if (sumVal != null) {
response = AggregateResponse.newBuilder().addFirstPart(
ci.getProtoForPromotedType(sumVal)).build();
}
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
log.debug("Sum from this region is "
+ env.getRegion().getRegionNameAsString() + ": " + sum);
done.run(response);
}
/**
* Gives the row count for the given column family and column qualifier, in
* the given row range as defined in the Scan object.
* @throws IOException
*/
@Override
public void getRowNum(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
AggregateResponse response = null;
long counter = 0l;
List<KeyValue> results = new ArrayList<KeyValue>();
InternalScanner scanner = null;
try {
Scan scan = ProtobufUtil.toScan(request.getScan());
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] qualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
qualifier = qualifiers.pollFirst();
}
if (scan.getFilter() == null && qualifier == null)
scan.setFilter(new FirstKeyOnlyFilter());
scanner = env.getRegion().getScanner(scan);
boolean hasMoreRows = false;
do {
hasMoreRows = scanner.next(results);
if (results.size() > 0) {
counter++;
}
results.clear();
} while (hasMoreRows);
ByteBuffer bb = ByteBuffer.allocate(8).putLong(counter);
bb.rewind();
response = AggregateResponse.newBuilder().addFirstPart(
ByteString.copyFrom(bb)).build();
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
log.info("Row counter from this region is "
+ env.getRegion().getRegionNameAsString() + ": " + counter);
done.run(response);
}
/**
* Gives a Pair with first object as Sum and second object as row count,
* computed for a given combination of column qualifier and column family in
* the given row range as defined in the Scan object. In its current
* implementation, it takes one column family and one column qualifier (if
* provided). In case of null column qualifier, an aggregate sum over all the
* entire column family will be returned.
* <p>
* The average is computed in
* {@link AggregationClient#avg(byte[], ColumnInterpreter, Scan)} by
* processing results from all regions, so its "ok" to pass sum and a Long
* type.
*/
@Override
public void getAvg(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
AggregateResponse response = null;
InternalScanner scanner = null;
try {
ColumnInterpreter<T, S> ci = constructColumnInterpreterFromRequest(request);
S sumVal = null;
Long rowCountVal = 0l;
Scan scan = ProtobufUtil.toScan(request.getScan());
scanner = env.getRegion().getScanner(scan);
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] qualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
qualifier = qualifiers.pollFirst();
}
List<KeyValue> results = new ArrayList<KeyValue>();
boolean hasMoreRows = false;
do {
results.clear();
hasMoreRows = scanner.next(results);
for (KeyValue kv : results) {
sumVal = ci.add(sumVal, ci.castToReturnType(ci.getValue(colFamily,
qualifier, kv)));
}
rowCountVal++;
} while (hasMoreRows);
if (sumVal != null) {
ByteString first = ci.getProtoForPromotedType(sumVal);
AggregateResponse.Builder pair = AggregateResponse.newBuilder();
pair.addFirstPart(first);
ByteBuffer bb = ByteBuffer.allocate(8).putLong(rowCountVal);
bb.rewind();
pair.setSecondPart(ByteString.copyFrom(bb));
response = pair.build();
}
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
done.run(response);
}
/**
* Gives a Pair with first object a List containing Sum and sum of squares,
* and the second object as row count. It is computed for a given combination of
* column qualifier and column family in the given row range as defined in the
* Scan object. In its current implementation, it takes one column family and
* one column qualifier (if provided). The idea is get the value of variance first:
* the average of the squares less the square of the average a standard
* deviation is square root of variance.
*/
@Override
public void getStd(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
InternalScanner scanner = null;
AggregateResponse response = null;
try {
ColumnInterpreter<T, S> ci = constructColumnInterpreterFromRequest(request);
S sumVal = null, sumSqVal = null, tempVal = null;
long rowCountVal = 0l;
Scan scan = ProtobufUtil.toScan(request.getScan());
scanner = env.getRegion().getScanner(scan);
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] qualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
qualifier = qualifiers.pollFirst();
}
List<KeyValue> results = new ArrayList<KeyValue>();
boolean hasMoreRows = false;
do {
tempVal = null;
hasMoreRows = scanner.next(results);
for (KeyValue kv : results) {
tempVal = ci.add(tempVal, ci.castToReturnType(ci.getValue(colFamily,
qualifier, kv)));
}
results.clear();
sumVal = ci.add(sumVal, tempVal);
sumSqVal = ci.add(sumSqVal, ci.multiply(tempVal, tempVal));
rowCountVal++;
} while (hasMoreRows);
if (sumVal != null) {
ByteString first_sumVal = ci.getProtoForPromotedType(sumVal);
ByteString first_sumSqVal = ci.getProtoForPromotedType(sumSqVal);
AggregateResponse.Builder pair = AggregateResponse.newBuilder();
pair.addFirstPart(first_sumVal);
pair.addFirstPart(first_sumSqVal);
ByteBuffer bb = ByteBuffer.allocate(8).putLong(rowCountVal);
bb.rewind();
pair.setSecondPart(ByteString.copyFrom(bb));
response = pair.build();
}
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
done.run(response);
}
/**
* Gives a List containing sum of values and sum of weights.
* It is computed for the combination of column
* family and column qualifier(s) in the given row range as defined in the
* Scan object. In its current implementation, it takes one column family and
* two column qualifiers. The first qualifier is for values column and
* the second qualifier (optional) is for weight column.
*/
@Override
public void getMedian(RpcController controller, AggregateArgument request,
RpcCallback<AggregateResponse> done) {
AggregateResponse response = null;
InternalScanner scanner = null;
try {
ColumnInterpreter<T, S> ci = constructColumnInterpreterFromRequest(request);
S sumVal = null, sumWeights = null, tempVal = null, tempWeight = null;
Scan scan = ProtobufUtil.toScan(request.getScan());
scanner = env.getRegion().getScanner(scan);
byte[] colFamily = scan.getFamilies()[0];
NavigableSet<byte[]> qualifiers = scan.getFamilyMap().get(colFamily);
byte[] valQualifier = null, weightQualifier = null;
if (qualifiers != null && !qualifiers.isEmpty()) {
valQualifier = qualifiers.pollFirst();
// if weighted median is requested, get qualifier for the weight column
weightQualifier = qualifiers.pollLast();
}
List<KeyValue> results = new ArrayList<KeyValue>();
boolean hasMoreRows = false;
do {
tempVal = null;
tempWeight = null;
hasMoreRows = scanner.next(results);
for (KeyValue kv : results) {
tempVal = ci.add(tempVal, ci.castToReturnType(ci.getValue(colFamily,
valQualifier, kv)));
if (weightQualifier != null) {
tempWeight = ci.add(tempWeight,
ci.castToReturnType(ci.getValue(colFamily, weightQualifier, kv)));
}
}
results.clear();
sumVal = ci.add(sumVal, tempVal);
sumWeights = ci.add(sumWeights, tempWeight);
} while (hasMoreRows);
ByteString first_sumVal = ci.getProtoForPromotedType(sumVal);
S s = sumWeights == null ? ci.castToReturnType(ci.getMinValue()) : sumWeights;
ByteString first_sumWeights = ci.getProtoForPromotedType(s);
AggregateResponse.Builder pair = AggregateResponse.newBuilder();
pair.addFirstPart(first_sumVal);
pair.addFirstPart(first_sumWeights);
response = pair.build();
} catch (IOException e) {
ResponseConverter.setControllerException(controller, e);
} finally {
if (scanner != null) {
try {
scanner.close();
} catch (IOException ignored) {}
}
}
done.run(response);
}
@SuppressWarnings("unchecked")
ColumnInterpreter<T,S> constructColumnInterpreterFromRequest(
AggregateArgument request) throws IOException {
String className = request.getInterpreterClassName();
Class<?> cls;
try {
cls = Class.forName(className);
ColumnInterpreter<T,S> ci = (ColumnInterpreter<T, S>) cls.newInstance();
if (request.hasInterpreterSpecificBytes()) {
ci.initialize(request.getInterpreterSpecificBytes());
}
return ci;
} catch (ClassNotFoundException e) {
throw new IOException(e);
} catch (InstantiationException e) {
throw new IOException(e);
} catch (IllegalAccessException e) {
throw new IOException(e);
}
}
@Override
public Service getService() {
return this;
}
/**
* Stores a reference to the coprocessor environment provided by the
* {@link org.apache.hadoop.hbase.regionserver.RegionCoprocessorHost} from the region where this
* coprocessor is loaded. Since this is a coprocessor endpoint, it always expects to be loaded
* on a table region, so always expects this to be an instance of
* {@link RegionCoprocessorEnvironment}.
* @param env the environment provided by the coprocessor host
* @throws IOException if the provided environment is not an instance of
* {@code RegionCoprocessorEnvironment}
*/
@Override
public void start(CoprocessorEnvironment env) throws IOException {
if (env instanceof RegionCoprocessorEnvironment) {
this.env = (RegionCoprocessorEnvironment)env;
} else {
throw new CoprocessorException("Must be loaded on a table region!");
}
}
@Override
public void stop(CoprocessorEnvironment env) throws IOException {
// nothing to do
}
}