hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (Jira)" <j...@apache.org>
Subject [jira] [Work logged] (HIVE-24012) Support for rewriting with materialized views containing grouping sets
Date Fri, 07 Aug 2020 21:01:00 GMT

     [ https://issues.apache.org/jira/browse/HIVE-24012?focusedWorklogId=468020&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-468020
]

ASF GitHub Bot logged work on HIVE-24012:
-----------------------------------------

                Author: ASF GitHub Bot
            Created on: 07/Aug/20 21:00
            Start Date: 07/Aug/20 21:00
    Worklog Time Spent: 10m 
      Work Description: vineetgarg02 commented on a change in pull request #1374:
URL: https://github.com/apache/hive/pull/1374#discussion_r467263500



##########
File path: ql/src/java/org/apache/hadoop/hive/ql/optimizer/calcite/rules/views/HiveMaterializedViewUtils.java
##########
@@ -0,0 +1,368 @@
+/*
+ * 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.optimizer.calcite.rules.views;
+
+import com.google.common.collect.ImmutableList;
+import java.math.BigDecimal;
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.List;
+import java.util.concurrent.TimeUnit;
+import org.apache.calcite.adapter.druid.DruidQuery;
+import org.apache.calcite.interpreter.BindableConvention;
+import org.apache.calcite.plan.RelOptCluster;
+import org.apache.calcite.plan.RelOptMaterialization;
+import org.apache.calcite.plan.RelOptUtil;
+import org.apache.calcite.plan.hep.HepPlanner;
+import org.apache.calcite.plan.hep.HepProgramBuilder;
+import org.apache.calcite.rel.RelNode;
+import org.apache.calcite.rel.RelVisitor;
+import org.apache.calcite.rel.core.Aggregate;
+import org.apache.calcite.rel.core.Aggregate.Group;
+import org.apache.calcite.rel.core.AggregateCall;
+import org.apache.calcite.rel.core.Filter;
+import org.apache.calcite.rel.core.Project;
+import org.apache.calcite.rel.core.TableScan;
+import org.apache.calcite.rex.RexBuilder;
+import org.apache.calcite.rex.RexInputRef;
+import org.apache.calcite.rex.RexNode;
+import org.apache.calcite.sql.fun.SqlStdOperatorTable;
+import org.apache.calcite.tools.RelBuilder;
+import org.apache.calcite.util.ImmutableBitSet;
+import org.apache.hadoop.hive.common.ValidTxnWriteIdList;
+import org.apache.hadoop.hive.common.ValidWriteIdList;
+import org.apache.hadoop.hive.conf.HiveConf;
+import org.apache.hadoop.hive.metastore.api.CreationMetadata;
+import org.apache.hadoop.hive.ql.lockmgr.LockException;
+import org.apache.hadoop.hive.ql.metadata.Table;
+import org.apache.hadoop.hive.ql.optimizer.calcite.HiveRelFactories;
+import org.apache.hadoop.hive.ql.optimizer.calcite.RelOptHiveTable;
+import org.apache.hadoop.hive.ql.optimizer.calcite.reloperators.HiveFilter;
+import org.apache.hadoop.hive.ql.optimizer.calcite.reloperators.HiveGroupingID;
+import org.apache.hadoop.hive.ql.optimizer.calcite.reloperators.HiveProject;
+import org.apache.hadoop.hive.ql.optimizer.calcite.reloperators.HiveRelNode;
+import org.apache.hadoop.hive.ql.optimizer.calcite.reloperators.HiveTableScan;
+import org.apache.hadoop.hive.ql.parse.DruidSqlOperatorConverter;
+import org.apache.hadoop.hive.ql.session.SessionState;
+import org.apache.hive.common.util.TxnIdUtils;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+
+import static org.apache.hadoop.hive.conf.Constants.MATERIALIZED_VIEW_REWRITING_TIME_WINDOW;
+
+public class HiveMaterializedViewUtils {
+
+  private static final Logger LOG = LoggerFactory.getLogger(HiveMaterializedViewUtils.class);
+
+
+  private HiveMaterializedViewUtils() {}
+
+  public static Table extractTable(RelOptMaterialization materialization) {
+    RelOptHiveTable cachedMaterializedViewTable;
+    if (materialization.tableRel instanceof Project) {
+      // There is a Project on top (due to nullability)
+      cachedMaterializedViewTable = (RelOptHiveTable) materialization.tableRel.getInput(0).getTable();
+    } else {
+      cachedMaterializedViewTable = (RelOptHiveTable) materialization.tableRel.getTable();
+    }
+    return cachedMaterializedViewTable.getHiveTableMD();
+  }
+
+  /**
+   * Utility method that returns whether a materialized view is outdated (true), not outdated
+   * (false), or it cannot be determined (null). The latest case may happen e.g. when the
+   * materialized view definition uses external tables.
+   */
+  public static Boolean isOutdatedMaterializedView(Table materializedViewTable, final ValidTxnWriteIdList
currentTxnWriteIds,
+      long defaultTimeWindow, List<String> tablesUsed, boolean forceMVContentsUpToDate)
{
+    // Check if materialization defined its own invalidation time window
+    String timeWindowString = materializedViewTable.getProperty(MATERIALIZED_VIEW_REWRITING_TIME_WINDOW);
+    long timeWindow = org.apache.commons.lang3.StringUtils.isEmpty(timeWindowString) ? defaultTimeWindow
:
+        HiveConf.toTime(timeWindowString,
+            HiveConf.getDefaultTimeUnit(HiveConf.ConfVars.HIVE_MATERIALIZED_VIEW_REWRITING_TIME_WINDOW),
+            TimeUnit.MILLISECONDS);
+    CreationMetadata creationMetadata = materializedViewTable.getCreationMetadata();
+    boolean outdated = false;
+    if (timeWindow < 0L) {
+      // We only consider the materialized view to be outdated if forceOutdated = true, i.e.,
+      // if it is a rebuild. Otherwise, it passed the test and we use it as it is.
+      outdated = forceMVContentsUpToDate;
+    } else {
+      // Check whether the materialized view is invalidated
+      if (forceMVContentsUpToDate || timeWindow == 0L || creationMetadata.getMaterializationTime()
< System.currentTimeMillis() - timeWindow) {
+        if (currentTxnWriteIds == null) {
+          LOG.debug("Materialized view " + materializedViewTable.getFullyQualifiedName()
+
+              " ignored for rewriting as we could not obtain current txn ids");
+          return null;
+        }
+        if (creationMetadata.getValidTxnList() == null ||
+            creationMetadata.getValidTxnList().isEmpty()) {
+          LOG.debug("Materialized view " + materializedViewTable.getFullyQualifiedName()
+
+              " ignored for rewriting as we could not obtain materialization txn ids");
+          return null;
+        }
+        boolean ignore = false;
+        ValidTxnWriteIdList mvTxnWriteIds = new ValidTxnWriteIdList(
+            creationMetadata.getValidTxnList());
+        for (String qName : tablesUsed) {
+          // Note. If the materialized view does not contain a table that is contained in
the query,
+          // we do not need to check whether that specific table is outdated or not. If a
rewriting
+          // is produced in those cases, it is because that additional table is joined with
the
+          // existing tables with an append-columns only join, i.e., PK-FK + not null.
+          if (!creationMetadata.getTablesUsed().contains(qName)) {
+            continue;
+          }
+          ValidWriteIdList tableCurrentWriteIds = currentTxnWriteIds.getTableValidWriteIdList(qName);
+          if (tableCurrentWriteIds == null) {
+            // Uses non-transactional table, cannot be considered
+            LOG.debug("Materialized view " + materializedViewTable.getFullyQualifiedName()
+
+                " ignored for rewriting as it is outdated and cannot be considered for "
+
+                " rewriting because it uses non-transactional table " + qName);
+            ignore = true;
+            break;
+          }
+          ValidWriteIdList tableWriteIds = mvTxnWriteIds.getTableValidWriteIdList(qName);
+          if (tableWriteIds == null) {
+            // This should not happen, but we ignore for safety
+            LOG.warn("Materialized view " + materializedViewTable.getFullyQualifiedName()
+
+                " ignored for rewriting as details about txn ids for table " + qName +
+                " could not be found in " + mvTxnWriteIds);
+            ignore = true;
+            break;
+          }
+          if (!outdated && !TxnIdUtils.checkEquivalentWriteIds(tableCurrentWriteIds,
tableWriteIds)) {
+            LOG.debug("Materialized view " + materializedViewTable.getFullyQualifiedName()
+
+                " contents are outdated");
+            outdated = true;
+          }
+        }
+        if (ignore) {
+          return null;
+        }
+      }
+    }
+    return outdated;
+  }
+
+  /**
+   * Method to enrich the materialization query contained in the input with
+   * its invalidation.
+   */
+  public static RelOptMaterialization augmentMaterializationWithTimeInformation(
+      RelOptMaterialization materialization, String validTxnsList,
+      ValidTxnWriteIdList materializationTxnList) throws LockException {
+    // Extract tables used by the query which will in turn be used to generate
+    // the corresponding txn write ids
+    List<String> tablesUsed = new ArrayList<>();
+    new RelVisitor() {
+      @Override
+      public void visit(RelNode node, int ordinal, RelNode parent) {
+        if (node instanceof TableScan) {
+          TableScan ts = (TableScan) node;
+          tablesUsed.add(((RelOptHiveTable) ts.getTable()).getHiveTableMD().getFullyQualifiedName());
+        }
+        super.visit(node, ordinal, parent);
+      }
+    }.go(materialization.queryRel);
+    ValidTxnWriteIdList currentTxnList =
+        SessionState.get().getTxnMgr().getValidWriteIds(tablesUsed, validTxnsList);
+    // Augment
+    final RexBuilder rexBuilder = materialization.queryRel.getCluster().getRexBuilder();
+    final HepProgramBuilder augmentMaterializationProgram = new HepProgramBuilder()
+        .addRuleInstance(new HiveAugmentMaterializationRule(rexBuilder, currentTxnList, materializationTxnList));
+    final HepPlanner augmentMaterializationPlanner = new HepPlanner(
+        augmentMaterializationProgram.build());
+    augmentMaterializationPlanner.setRoot(materialization.queryRel);
+    final RelNode modifiedQueryRel = augmentMaterializationPlanner.findBestExp();
+    return new RelOptMaterialization(materialization.tableRel, modifiedQueryRel,
+        null, materialization.qualifiedTableName);
+  }
+
+  /**
+   * If a materialization does not contain grouping sets, it returns the materialization
+   * itself. Otherwise, it will create one materialization for each grouping set.
+   * For each grouping set, the query for the materialization will consist of the group by
+   * columns in the grouping set, followed by a projection to recreate the literal null
+   * values. The view scan will consist of the scan over the materialization followed by
a
+   * filter on the grouping id value corresponding to that grouping set.
+   */
+  public static List<RelOptMaterialization> deriveGroupingSetsMaterializedViews(RelOptMaterialization
materialization) {
+    final RelNode query = materialization.queryRel;
+    final Project project;
+    final Aggregate aggregate;
+    if (query instanceof Aggregate) {
+      project = null;
+      aggregate = (Aggregate) query;
+    } else if (query instanceof Project && query.getInput(0) instanceof Aggregate)
{
+      project = (Project) query;
+      aggregate = (Aggregate) query.getInput(0);
+    } else {
+      project = null;
+      aggregate = null;
+    }
+    if (aggregate == null) {
+      // Not an aggregate materialized view, return original materialization
+      return Collections.singletonList(materialization);
+    }
+    if (aggregate.getGroupType() == Group.SIMPLE) {
+      // Not a grouping sets materialized view, return original materialization
+      return Collections.singletonList(materialization);
+    }
+    int groupingIdIndex = -1;
+    for (int i = 0; i < aggregate.getAggCallList().size(); i++) {
+      if (aggregate.getAggCallList().get(i).getAggregation() == HiveGroupingID.INSTANCE)
{
+        groupingIdIndex = aggregate.getGroupCount() + i;
+        break;
+      }
+    }
+    assert groupingIdIndex != -1;

Review comment:
       Precondition?




----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


Issue Time Tracking
-------------------

    Worklog Id:     (was: 468020)
    Time Spent: 20m  (was: 10m)

> Support for rewriting with materialized views containing grouping sets
> ----------------------------------------------------------------------
>
>                 Key: HIVE-24012
>                 URL: https://issues.apache.org/jira/browse/HIVE-24012
>             Project: Hive
>          Issue Type: Sub-task
>          Components: Materialized views
>            Reporter: Jesus Camacho Rodriguez
>            Assignee: Jesus Camacho Rodriguez
>            Priority: Major
>              Labels: pull-request-available
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Rewriting is not triggered for materialized views containing grouping sets. This issue
implements an extension from Hive side to trigger additional rewritings for materialized views
containing grouping sets.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Mime
View raw message