You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "ASF GitHub Bot (Jira)" <ji...@apache.org> on 2022/05/31 09:56:00 UTC
[jira] [Work logged] (HIVE-26243) Add vectorized implementation of the 'ds_kll_sketch' UDAF
[ https://issues.apache.org/jira/browse/HIVE-26243?focusedWorklogId=776193&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-776193 ]
ASF GitHub Bot logged work on HIVE-26243:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 31/May/22 09:55
Start Date: 31/May/22 09:55
Worklog Time Spent: 10m
Work Description: kgyrtkirk commented on code in PR #3317:
URL: https://github.com/apache/hive/pull/3317#discussion_r885430661
##########
standalone-metastore/pom.xml:
##########
@@ -181,6 +182,17 @@
<artifactId>commons-lang3</artifactId>
<version>${commons-lang3.version}</version>
</dependency>
+ <dependency>
+ <groupId>org.apache.datasketches</groupId>
Review Comment:
I think this has nothing to do with this patch
##########
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/common/histogram/kll/KllUtils.java:
##########
@@ -0,0 +1,106 @@
+/*
+ * 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.hive.common.histogram.kll;
+
+import org.apache.datasketches.kll.KllFloatsSketch;
+import org.apache.datasketches.memory.Memory;
+import org.apache.hadoop.hive.ql.util.JavaDataModel;
+
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
+
+/**
+ * KLL serialization utilities.
+ */
+public class KllUtils {
Review Comment:
unrelated to changes?
##########
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/common/histogram/KllHistogramEstimator.java:
##########
@@ -0,0 +1,83 @@
+/*
+ * 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.hive.common.histogram;
+
+import org.apache.datasketches.kll.KllFloatsSketch;
+import org.apache.hadoop.hive.common.histogram.kll.KllUtils;
+import org.apache.hadoop.hive.common.type.HiveDecimal;
+import org.apache.hadoop.hive.ql.util.JavaDataModel;
+
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+
+public class KllHistogramEstimator {
Review Comment:
remove these classes from the metastore...
##########
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/common/histogram/kll/KllUtils.java:
##########
@@ -0,0 +1,106 @@
+/*
+ * 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.hive.common.histogram.kll;
+
+import org.apache.datasketches.kll.KllFloatsSketch;
+import org.apache.datasketches.memory.Memory;
+import org.apache.hadoop.hive.ql.util.JavaDataModel;
+
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.OutputStream;
+
+/**
+ * KLL serialization utilities.
+ */
+public class KllUtils {
+
+ private KllUtils() {
+ throw new AssertionError("Suppress default constructor for non instantiation");
+ }
+
+ /**
+ * KLL is serialized according to what provided by data-sketches library
+ * @param out output stream to write to
+ * @param kll KLL sketch that needs to be serialized
+ * @throws IOException if an error occurs during serialization
+ */
+ public static void serializeKll(OutputStream out, KllFloatsSketch kll) throws IOException {
+ out.write(kll.toByteArray());
+ }
Review Comment:
do we really need these methods/classes/etc?
this really just binds the class name; the method name with no good reason...no interface; what's the architectural value here I don't see?
##########
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/common/histogram/KllHistogramEstimatorFactory.java:
##########
@@ -0,0 +1,75 @@
+/*
+ * 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.hive.common.histogram;
+
+import org.apache.hadoop.hive.common.histogram.kll.KllUtils;
+
+public class KllHistogramEstimatorFactory {
+
+ private KllHistogramEstimatorFactory() {
+ throw new AssertionError("Suppress default constructor for non instantiation");
+ }
+
+ /**
+ * This function deserializes the serialized KLL histogram estimator from a byte array.
+ * @param buf to deserialize
+ * @return KLL histogram estimator
+ */
+ public static KllHistogramEstimator getKllHistogramEstimator(byte[] buf) {
+ return new KllHistogramEstimator(KllUtils.deserializeKll(buf, 0, buf.length));
+ }
+
+ /**
+ * This function deserializes the serialized KLL histogram estimator from a byte array.
+ * @param buf to deserialize
+ * @param start start index for deserialization
+ * @param len start+len is deserialized
+ * @return KLL histogram estimator
+ */
+ public static KllHistogramEstimator getKllHistogramEstimator(byte[] buf, int start, int len) {
+ return new KllHistogramEstimator(KllUtils.deserializeKll(buf, start, len));
+ }
+
+ /**
+ * This method creates an empty histogram estimator with a KLL sketch with k=200.
+ * @return an empty histogram estimator with a KLL sketch with k=200
+ */
+ public static KllHistogramEstimator getEmptyHistogramEstimator() {
Review Comment:
there are 3 `getEmptyHistogramEstimator` -s from which only 2 is being used; and I think 1 will be too much already...do we need these at all? what's the point of introducing these???
##########
ql/src/gen/vectorization/UDAFTemplates/VectorUDAFComputeKLL.txt:
##########
@@ -0,0 +1,307 @@
+/*
+ * 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.hive.ql.exec.vector.expressions.aggregates;
+
+import org.apache.hadoop.hive.common.histogram.KllHistogramEstimator;
+import org.apache.hadoop.hive.common.histogram.KllHistogramEstimatorFactory;
+import org.apache.hadoop.hive.ql.exec.Description;
+import org.apache.hadoop.hive.ql.exec.vector.BytesColumnVector;
+import org.apache.hadoop.hive.ql.exec.vector.ColumnVector;
+#IF COMPLETE
+import org.apache.hadoop.hive.ql.exec.vector.<InputColumnVectorType>;
+#ENDIF COMPLETE
+import org.apache.hadoop.hive.ql.exec.vector.VectorAggregationBufferRow;
+import org.apache.hadoop.hive.ql.exec.vector.VectorAggregationDesc;
+import org.apache.hadoop.hive.ql.exec.vector.VectorizedRowBatch;
+import org.apache.hadoop.hive.ql.metadata.HiveException;
+import org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator;
+import org.apache.hadoop.hive.ql.util.JavaDataModel;
+
+/**
+ * Generated from template VectorUDAFComputeKLL.txt.
Review Comment:
this seems to be the vectorized pair of `ds_kll_sketch` or not?
can we keep the naming conventions? it will be hard to follow what-is-what....without doing so
##########
ql/src/gen/vectorization/UDAFTemplates/VectorUDAFComputeKLL.txt:
##########
@@ -0,0 +1,307 @@
+/*
+ * 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.hive.ql.exec.vector.expressions.aggregates;
+
+import org.apache.hadoop.hive.common.histogram.KllHistogramEstimator;
+import org.apache.hadoop.hive.common.histogram.KllHistogramEstimatorFactory;
+import org.apache.hadoop.hive.ql.exec.Description;
+import org.apache.hadoop.hive.ql.exec.vector.BytesColumnVector;
+import org.apache.hadoop.hive.ql.exec.vector.ColumnVector;
+#IF COMPLETE
+import org.apache.hadoop.hive.ql.exec.vector.<InputColumnVectorType>;
+#ENDIF COMPLETE
+import org.apache.hadoop.hive.ql.exec.vector.VectorAggregationBufferRow;
+import org.apache.hadoop.hive.ql.exec.vector.VectorAggregationDesc;
+import org.apache.hadoop.hive.ql.exec.vector.VectorizedRowBatch;
+import org.apache.hadoop.hive.ql.metadata.HiveException;
+import org.apache.hadoop.hive.ql.udf.generic.GenericUDAFEvaluator;
+import org.apache.hadoop.hive.ql.util.JavaDataModel;
+
+/**
+ * Generated from template VectorUDAFComputeKLL.txt.
+ */
+@Description(name = "ds_kll_sketch", value = "_FUNC_(x) "
+ + "Returns a KllFloatsSketch in a serialized form as a binary blob."
+ + " Values must be of type float.")
+public class <ClassName> extends VectorAggregateExpression {
+
+ public <ClassName>() {
+ super();
+ }
+
+ public <ClassName>(VectorAggregationDesc vecAggrDesc) {
+ super(vecAggrDesc);
+ }
+
+ @Override
+ public AggregationBuffer getNewAggregationBuffer() throws HiveException {
+ return new Aggregation();
+ }
+
+ @Override
+ public void aggregateInput(AggregationBuffer agg, VectorizedRowBatch batch) throws HiveException {
+ inputExpression.evaluate(batch);
+
+#IF COMPLETE
+ <InputColumnVectorType> inputColumn = (<InputColumnVectorType>) batch.cols[this.inputExpression.getOutputColumnNum()];
+#ENDIF COMPLETE
+#IF MERGING
+ BytesColumnVector inputColumn = (BytesColumnVector) batch.cols[this.inputExpression.getOutputColumnNum()];
+#ENDIF MERGING
+
+ int batchSize = batch.size;
+
+ if (batchSize == 0) {
+ return;
+ }
+
+ Aggregation myagg = (Aggregation) agg;
+
+#IF COMPLETE
+ myagg.prepare();
+ if (inputColumn.noNulls) {
+ if (inputColumn.isRepeating) {
+ for (int i = 0; i < batchSize; i++) {
+ myagg.estimator.addToEstimator(inputColumn.vector[0]);
+ }
+ } else {
+ if (batch.selectedInUse) {
+ for (int s = 0; s < batchSize; s++) {
+ int i = batch.selected[s];
+ myagg.estimator.addToEstimator(inputColumn.vector[i]);
+ }
+ } else {
+ for (int i = 0; i < batchSize; i++) {
+ myagg.estimator.addToEstimator(inputColumn.vector[i]);
+ }
+ }
+ }
+ } else {
+ if (inputColumn.isRepeating) {
+ if (!inputColumn.isNull[0]) {
+ for (int i = 0; i < batchSize; i++) {
+ myagg.estimator.addToEstimator(inputColumn.vector[0]);
+ }
+ }
+ } else {
+ if (batch.selectedInUse) {
+ for (int j = 0; j < batchSize; ++j) {
+ int i = batch.selected[j];
+ if (!inputColumn.isNull[i]) {
+ myagg.estimator.addToEstimator(inputColumn.vector[i]);
+ }
+ }
+ } else {
+ for (int i = 0; i < batchSize; i++) {
+ if (!inputColumn.isNull[i]) {
+ myagg.estimator.addToEstimator(inputColumn.vector[i]);
+ }
+ }
+ }
+ }
+ }
+#ENDIF COMPLETE
+#IF MERGING
+ if (inputColumn.isRepeating) {
+ if (!inputColumn.isNull[0] && inputColumn.length[0] > 0) {
+ myagg.prepare();
+ KllHistogramEstimator mergingKLL = KllHistogramEstimatorFactory.getKllHistogramEstimator(
Review Comment:
do we really have to create and bind 1-by-1 vectorized implementations for the datasketches functions ?
function classes kinda do the same stuff... like `DATA_TO_SKETCH` ;
Issue Time Tracking
-------------------
Worklog Id: (was: 776193)
Remaining Estimate: 0h
Time Spent: 10m
> Add vectorized implementation of the 'ds_kll_sketch' UDAF
> ---------------------------------------------------------
>
> Key: HIVE-26243
> URL: https://issues.apache.org/jira/browse/HIVE-26243
> Project: Hive
> Issue Type: Improvement
> Components: UDF, Vectorization
> Affects Versions: 4.0.0-alpha-2
> Reporter: Alessandro Solimando
> Assignee: Alessandro Solimando
> Priority: Major
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> _ds_kll_sketch_ UDAF does not have a vectorized implementation at the moment, the present ticket aims at bridging this gap.
> This is particularly important because vectorization has an "all or nothing" approach, so if this function is used at the side of vectorized functions, they won't be able to benefit from vectorized execution.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)