vdiravka commented on a change in pull request #1466: DRILL-6381: Add support for index based planning and execution URL: https://github.com/apache/drill/pull/1466#discussion_r223670746
########## File path: exec/java-exec/src/main/java/org/apache/drill/exec/planner/index/IndexPlanUtils.java ########## @@ -0,0 +1,872 @@ +/* + * 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.drill.exec.planner.index; + +import java.util.ArrayList; +import java.util.List; +import java.util.Map; +import java.util.Set; + +import org.apache.drill.shaded.guava.com.google.common.collect.ImmutableList; +import org.apache.drill.shaded.guava.com.google.common.collect.Lists; +import org.apache.drill.shaded.guava.com.google.common.collect.Maps; +import org.apache.drill.shaded.guava.com.google.common.collect.Sets; + +import org.apache.calcite.plan.RelTraitSet; +import org.apache.calcite.plan.volcano.RelSubset; +import org.apache.calcite.rel.RelCollation; +import org.apache.calcite.rel.RelCollationTraitDef; +import org.apache.calcite.rel.RelCollations; +import org.apache.calcite.rel.RelFieldCollation; +import org.apache.calcite.rel.RelNode; +import org.apache.calcite.rel.core.Sort; +import org.apache.calcite.rex.RexBuilder; +import org.apache.calcite.rex.RexUtil; +import org.apache.calcite.rex.RexLiteral; +import org.apache.calcite.sql.SqlKind; +import org.apache.drill.common.expression.FieldReference; +import org.apache.drill.common.expression.LogicalExpression; +import org.apache.drill.common.expression.SchemaPath; +import org.apache.drill.exec.physical.base.DbGroupScan; +import org.apache.drill.exec.physical.base.GroupScan; +import org.apache.drill.exec.physical.base.IndexGroupScan; +import org.apache.drill.exec.planner.common.DrillProjectRelBase; +import org.apache.drill.exec.planner.common.DrillScanRelBase; +import org.apache.drill.exec.planner.fragment.DistributionAffinity; +import org.apache.drill.exec.planner.logical.DrillOptiq; +import org.apache.drill.exec.planner.logical.DrillParseContext; +import org.apache.drill.exec.planner.logical.DrillScanRel; +import org.apache.drill.exec.planner.physical.DrillDistributionTrait; +import org.apache.drill.exec.planner.physical.Prel; +import org.apache.drill.exec.planner.physical.PrelUtil; +import org.apache.drill.exec.planner.physical.ScanPrel; +import org.apache.drill.exec.planner.physical.ProjectPrel; +import org.apache.drill.exec.planner.common.OrderedRel; +import org.apache.calcite.rel.type.RelDataType; +import org.apache.calcite.rel.type.RelDataTypeField; +import org.apache.calcite.rex.RexInputRef; +import org.apache.calcite.rex.RexNode; + +public class IndexPlanUtils { + + public enum ConditionIndexed { + NONE, + PARTIAL, + FULL} + + /** + * Check if any of the fields of the index are present in a list of LogicalExpressions supplied + * as part of IndexableExprMarker + * @param exprMarker, the marker that has analyzed original index condition on top of original scan + * @param indexDesc + * @return ConditionIndexed.FULL, PARTIAL or NONE depending on whether all, some or no columns + * of the indexDesc are present in the list of LogicalExpressions supplied as part of exprMarker + * + */ + static public ConditionIndexed conditionIndexed(IndexableExprMarker exprMarker, IndexDescriptor indexDesc) { + Map<RexNode, LogicalExpression> mapRexExpr = exprMarker.getIndexableExpression(); + List<LogicalExpression> infoCols = Lists.newArrayList(); + infoCols.addAll(mapRexExpr.values()); + if (indexDesc.allColumnsIndexed(infoCols)) { + return ConditionIndexed.FULL; + } else if (indexDesc.someColumnsIndexed(infoCols)) { + return ConditionIndexed.PARTIAL; + } else { + return ConditionIndexed.NONE; + } + } + + /** + * check if we want to apply index rules on this scan, + * if group scan is not instance of DbGroupScan, or this DbGroupScan instance does not support secondary index, or + * this scan is already an index scan or Restricted Scan, do not apply index plan rules on it. + * @param scanRel + * @return + */ + static public boolean checkScan(DrillScanRel scanRel) { + GroupScan groupScan = scanRel.getGroupScan(); + if (groupScan instanceof DbGroupScan) { + DbGroupScan dbscan = ((DbGroupScan) groupScan); + //if we already applied index convert rule, and this scan is indexScan or restricted scan already, + //no more trying index convert rule + return dbscan.supportsSecondaryIndex() && (!dbscan.isIndexScan()) && (!dbscan.isRestrictedScan()); + } + return false; + } + + /** + * For a particular table scan for table T1 and an index on that table, find out if it is a covering index + * @return + */ + static public boolean isCoveringIndex(IndexCallContext indexContext, FunctionalIndexInfo functionInfo) { + if(functionInfo.hasFunctional()) { + //need info from full query + return queryCoveredByIndex(indexContext, functionInfo); + } + DbGroupScan groupScan = (DbGroupScan) getGroupScan(indexContext.getScan()); + List<LogicalExpression> tableCols = Lists.newArrayList(); + tableCols.addAll(groupScan.getColumns()); + return functionInfo.getIndexDesc().isCoveringIndex(tableCols); + } + + + /** + * This method is called only when the index has at least one functional indexed field. If there is no function field, + * we don't need to worry whether there could be paths not found in Scan. + * In functional case, we have to check all available (if needed) operators to find out if the query is covered or not. + * E.g. cast(a.b as INT) in project, a.b in Scan's rowType or columns, and cast(a.b as INT) + * is an indexed field named '$0'. In this case, by looking at Scan, we see only 'a.b' which is not in index. We have to + * look into Project, and if we see 'a.b' is only used in functional index expression cast(a.b as INT), then we know + * this Project+Scan is covered. + * @param indexContext + * @param functionInfo + * @return false if the query could not be covered by the index (should not create covering index plan) + */ + static private boolean queryCoveredByIndex(IndexCallContext indexContext, + FunctionalIndexInfo functionInfo) { + //for indexed functions, if relevant schemapaths are included in index(in indexed fields or non-indexed fields), + // check covering based on the local information we have: + // if references to schema paths in functional indexes disappear beyond capProject + + if (indexContext.getFilter() != null && indexContext.getUpperProject() == null) { + if( !isFullQuery(indexContext)) { Review comment: formatting ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
