use of org.apache.calcite.plan.RelOptMaterialization in project hive by apache.
the class HiveMaterializedViewFilterScanRule method apply.
protected void apply(RelOptRuleCall call, Project project, Filter filter, TableScan scan) {
RelOptPlanner planner = call.getPlanner();
List<RelOptMaterialization> materializations = (planner instanceof VolcanoPlanner) ? ((VolcanoPlanner) planner).getMaterializations() : ImmutableList.<RelOptMaterialization>of();
if (!materializations.isEmpty()) {
RelNode root = project.copy(project.getTraitSet(), Collections.singletonList(filter.copy(filter.getTraitSet(), Collections.singletonList((RelNode) scan))));
// Costing is done in transformTo(), so we call it repeatedly with all applicable
// materialized views and cheapest one will be picked
List<RelOptMaterialization> applicableMaterializations = VolcanoPlanner.getApplicableMaterializations(root, materializations);
for (RelOptMaterialization materialization : applicableMaterializations) {
List<RelNode> subs = new MaterializedViewSubstitutionVisitor(materialization.queryRel, root, relBuilderFactory).go(materialization.tableRel);
for (RelNode s : subs) {
call.transformTo(s);
}
}
}
}
use of org.apache.calcite.plan.RelOptMaterialization in project hive by apache.
the class Hive method getRewritingMaterializedViews.
/**
* Get the materialized views that have been enabled for rewriting from the
* metastore. If the materialized view is in the cache, we do not need to
* parse it to generate a logical plan for the rewriting. Instead, we
* return the version present in the cache.
*
* @return the list of materialized views available for rewriting
* @throws HiveException
*/
public List<RelOptMaterialization> getRewritingMaterializedViews() throws HiveException {
try {
// Final result
List<RelOptMaterialization> result = new ArrayList<>();
for (String dbName : getMSC().getAllDatabases()) {
// From metastore (for security)
List<String> tables = getMSC().getAllTables(dbName);
// Cached views (includes all)
Collection<RelOptMaterialization> cachedViews = HiveMaterializedViewsRegistry.get().getRewritingMaterializedViews(dbName);
if (cachedViews.isEmpty()) {
// Bail out: empty list
continue;
}
Map<String, RelOptMaterialization> qualifiedNameToView = new HashMap<String, RelOptMaterialization>();
for (RelOptMaterialization materialization : cachedViews) {
qualifiedNameToView.put(materialization.table.getQualifiedName().get(0), materialization);
}
for (String table : tables) {
// Compose qualified name
String fullyQualifiedName = dbName;
if (fullyQualifiedName != null && !fullyQualifiedName.isEmpty()) {
fullyQualifiedName = fullyQualifiedName + "." + table;
} else {
fullyQualifiedName = table;
}
RelOptMaterialization materialization = qualifiedNameToView.get(fullyQualifiedName);
if (materialization != null) {
// Add to final result set
result.add(materialization);
}
}
}
return result;
} catch (Exception e) {
throw new HiveException(e);
}
}
use of org.apache.calcite.plan.RelOptMaterialization in project calcite by apache.
the class AbstractMaterializedViewRule method perform.
/**
* Rewriting logic is based on "Optimizing Queries Using Materialized Views:
* A Practical, Scalable Solution" by Goldstein and Larson.
*
* <p>On the query side, rules matches a Project-node chain or node, where node
* is either an Aggregate or a Join. Subplan rooted at the node operator must
* be composed of one or more of the following operators: TableScan, Project,
* Filter, and Join.
*
* <p>For each join MV, we need to check the following:
* <ol>
* <li> The plan rooted at the Join operator in the view produces all rows
* needed by the plan rooted at the Join operator in the query.</li>
* <li> All columns required by compensating predicates, i.e., predicates that
* need to be enforced over the view, are available at the view output.</li>
* <li> All output expressions can be computed from the output of the view.</li>
* <li> All output rows occur with the correct duplication factor. We might
* rely on existing Unique-Key - Foreign-Key relationships to extract that
* information.</li>
* </ol>
*
* <p>In turn, for each aggregate MV, we need to check the following:
* <ol>
* <li> The plan rooted at the Aggregate operator in the view produces all rows
* needed by the plan rooted at the Aggregate operator in the query.</li>
* <li> All columns required by compensating predicates, i.e., predicates that
* need to be enforced over the view, are available at the view output.</li>
* <li> The grouping columns in the query are a subset of the grouping columns
* in the view.</li>
* <li> All columns required to perform further grouping are available in the
* view output.</li>
* <li> All columns required to compute output expressions are available in the
* view output.</li>
* </ol>
*
* <p>The rule contains multiple extensions compared to the original paper. One of
* them is the possibility of creating rewritings using Union operators, e.g., if
* the result of a query is partially contained in the materialized view.
*/
protected void perform(RelOptRuleCall call, Project topProject, RelNode node) {
final RexBuilder rexBuilder = node.getCluster().getRexBuilder();
final RelMetadataQuery mq = RelMetadataQuery.instance();
final RelOptPlanner planner = call.getPlanner();
final RexExecutor executor = Util.first(planner.getExecutor(), RexUtil.EXECUTOR);
final RelOptPredicateList predicates = RelOptPredicateList.EMPTY;
final RexSimplify simplify = new RexSimplify(rexBuilder, predicates, true, executor);
final List<RelOptMaterialization> materializations = (planner instanceof VolcanoPlanner) ? ((VolcanoPlanner) planner).getMaterializations() : ImmutableList.<RelOptMaterialization>of();
if (!materializations.isEmpty()) {
// try to generate a rewriting are met
if (!isValidPlan(topProject, node, mq)) {
return;
}
// Obtain applicable (filtered) materializations
// TODO: Filtering of relevant materializations needs to be
// improved so we gather only materializations that might
// actually generate a valid rewriting.
final List<RelOptMaterialization> applicableMaterializations = RelOptMaterializations.getApplicableMaterializations(node, materializations);
if (!applicableMaterializations.isEmpty()) {
// 2. Initialize all query related auxiliary data structures
// that will be used throughout query rewriting process
// Generate query table references
final Set<RelTableRef> queryTableRefs = mq.getTableReferences(node);
if (queryTableRefs == null) {
// Bail out
return;
}
// Extract query predicates
final RelOptPredicateList queryPredicateList = mq.getAllPredicates(node);
if (queryPredicateList == null) {
// Bail out
return;
}
final RexNode pred = simplify.simplify(RexUtil.composeConjunction(rexBuilder, queryPredicateList.pulledUpPredicates, false));
final Triple<RexNode, RexNode, RexNode> queryPreds = splitPredicates(rexBuilder, pred);
// Extract query equivalence classes. An equivalence class is a set
// of columns in the query output that are known to be equal.
final EquivalenceClasses qEC = new EquivalenceClasses();
for (RexNode conj : RelOptUtil.conjunctions(queryPreds.getLeft())) {
assert conj.isA(SqlKind.EQUALS);
RexCall equiCond = (RexCall) conj;
qEC.addEquivalenceClass((RexTableInputRef) equiCond.getOperands().get(0), (RexTableInputRef) equiCond.getOperands().get(1));
}
// rewrite the given query
for (RelOptMaterialization materialization : applicableMaterializations) {
RelNode view = materialization.tableRel;
Project topViewProject;
RelNode viewNode;
if (materialization.queryRel instanceof Project) {
topViewProject = (Project) materialization.queryRel;
viewNode = topViewProject.getInput();
} else {
topViewProject = null;
viewNode = materialization.queryRel;
}
// 3.1. View checks before proceeding
if (!isValidPlan(topViewProject, viewNode, mq)) {
// Skip it
continue;
}
// 3.2. Initialize all query related auxiliary data structures
// that will be used throughout query rewriting process
// Extract view predicates
final RelOptPredicateList viewPredicateList = mq.getAllPredicates(viewNode);
if (viewPredicateList == null) {
// Skip it
continue;
}
final RexNode viewPred = simplify.simplify(RexUtil.composeConjunction(rexBuilder, viewPredicateList.pulledUpPredicates, false));
final Triple<RexNode, RexNode, RexNode> viewPreds = splitPredicates(rexBuilder, viewPred);
// Extract view table references
final Set<RelTableRef> viewTableRefs = mq.getTableReferences(viewNode);
if (viewTableRefs == null) {
// Bail out
return;
}
// Extract view tables
MatchModality matchModality;
Multimap<RexTableInputRef, RexTableInputRef> compensationEquiColumns = ArrayListMultimap.create();
if (!queryTableRefs.equals(viewTableRefs)) {
// subset of query tables (add additional tables through joins if possible)
if (viewTableRefs.containsAll(queryTableRefs)) {
matchModality = MatchModality.QUERY_PARTIAL;
final EquivalenceClasses vEC = new EquivalenceClasses();
for (RexNode conj : RelOptUtil.conjunctions(viewPreds.getLeft())) {
assert conj.isA(SqlKind.EQUALS);
RexCall equiCond = (RexCall) conj;
vEC.addEquivalenceClass((RexTableInputRef) equiCond.getOperands().get(0), (RexTableInputRef) equiCond.getOperands().get(1));
}
if (!compensatePartial(viewTableRefs, vEC, queryTableRefs, compensationEquiColumns)) {
// Cannot rewrite, skip it
continue;
}
} else if (queryTableRefs.containsAll(viewTableRefs)) {
matchModality = MatchModality.VIEW_PARTIAL;
ViewPartialRewriting partialRewritingResult = compensateViewPartial(call.builder(), rexBuilder, mq, view, topProject, node, queryTableRefs, qEC, topViewProject, viewNode, viewTableRefs);
if (partialRewritingResult == null) {
// Cannot rewrite, skip it
continue;
}
// Rewrite succeeded
view = partialRewritingResult.newView;
topViewProject = partialRewritingResult.newTopViewProject;
viewNode = partialRewritingResult.newViewNode;
} else {
// Skip it
continue;
}
} else {
matchModality = MatchModality.COMPLETE;
}
// 4. We map every table in the query to a table with the same qualified
// name (all query tables are contained in the view, thus this is equivalent
// to mapping every table in the query to a view table).
final Multimap<RelTableRef, RelTableRef> multiMapTables = ArrayListMultimap.create();
for (RelTableRef queryTableRef1 : queryTableRefs) {
for (RelTableRef queryTableRef2 : queryTableRefs) {
if (queryTableRef1.getQualifiedName().equals(queryTableRef2.getQualifiedName())) {
multiMapTables.put(queryTableRef1, queryTableRef2);
}
}
}
// If a table is used multiple times, we will create multiple mappings,
// and we will try to rewrite the query using each of the mappings.
// Then, we will try to map every source table (query) to a target
// table (view), and if we are successful, we will try to create
// compensation predicates to filter the view results further
// (if needed).
final List<BiMap<RelTableRef, RelTableRef>> flatListMappings = generateTableMappings(multiMapTables);
for (BiMap<RelTableRef, RelTableRef> queryToViewTableMapping : flatListMappings) {
// TableMapping : mapping query tables -> view tables
// 4.0. If compensation equivalence classes exist, we need to add
// the mapping to the query mapping
final EquivalenceClasses currQEC = EquivalenceClasses.copy(qEC);
if (matchModality == MatchModality.QUERY_PARTIAL) {
for (Entry<RexTableInputRef, RexTableInputRef> e : compensationEquiColumns.entries()) {
// Copy origin
RelTableRef queryTableRef = queryToViewTableMapping.inverse().get(e.getKey().getTableRef());
RexTableInputRef queryColumnRef = RexTableInputRef.of(queryTableRef, e.getKey().getIndex(), e.getKey().getType());
// Add to query equivalence classes and table mapping
currQEC.addEquivalenceClass(queryColumnRef, e.getValue());
queryToViewTableMapping.put(e.getValue().getTableRef(), // identity
e.getValue().getTableRef());
}
}
// 4.1. Compute compensation predicates, i.e., predicates that need to be
// enforced over the view to retain query semantics. The resulting predicates
// are expressed using {@link RexTableInputRef} over the query.
// First, to establish relationship, we swap column references of the view
// predicates to point to query tables and compute equivalence classes.
final RexNode viewColumnsEquiPred = RexUtil.swapTableReferences(rexBuilder, viewPreds.getLeft(), queryToViewTableMapping.inverse());
final EquivalenceClasses queryBasedVEC = new EquivalenceClasses();
for (RexNode conj : RelOptUtil.conjunctions(viewColumnsEquiPred)) {
assert conj.isA(SqlKind.EQUALS);
RexCall equiCond = (RexCall) conj;
queryBasedVEC.addEquivalenceClass((RexTableInputRef) equiCond.getOperands().get(0), (RexTableInputRef) equiCond.getOperands().get(1));
}
Triple<RexNode, RexNode, RexNode> compensationPreds = computeCompensationPredicates(rexBuilder, simplify, currQEC, queryPreds, queryBasedVEC, viewPreds, queryToViewTableMapping);
if (compensationPreds == null && generateUnionRewriting) {
// Attempt partial rewriting using union operator. This rewriting
// will read some data from the view and the rest of the data from
// the query computation. The resulting predicates are expressed
// using {@link RexTableInputRef} over the view.
compensationPreds = computeCompensationPredicates(rexBuilder, simplify, queryBasedVEC, viewPreds, currQEC, queryPreds, queryToViewTableMapping.inverse());
if (compensationPreds == null) {
// This was our last chance to use the view, skip it
continue;
}
RexNode compensationColumnsEquiPred = compensationPreds.getLeft();
RexNode otherCompensationPred = RexUtil.composeConjunction(rexBuilder, ImmutableList.of(compensationPreds.getMiddle(), compensationPreds.getRight()), false);
assert !compensationColumnsEquiPred.isAlwaysTrue() || !otherCompensationPred.isAlwaysTrue();
// b. Generate union branch (query).
final RelNode unionInputQuery = rewriteQuery(call.builder(), rexBuilder, simplify, mq, compensationColumnsEquiPred, otherCompensationPred, topProject, node, queryToViewTableMapping, queryBasedVEC, currQEC);
if (unionInputQuery == null) {
// Skip it
continue;
}
// c. Generate union branch (view).
// We trigger the unifying method. This method will either create a Project
// or an Aggregate operator on top of the view. It will also compute the
// output expressions for the query.
final RelNode unionInputView = rewriteView(call.builder(), rexBuilder, simplify, mq, matchModality, true, view, topProject, node, topViewProject, viewNode, queryToViewTableMapping, currQEC);
if (unionInputView == null) {
// Skip it
continue;
}
// d. Generate final rewriting (union).
final RelNode result = createUnion(call.builder(), rexBuilder, topProject, unionInputQuery, unionInputView);
if (result == null) {
// Skip it
continue;
}
call.transformTo(result);
} else if (compensationPreds != null) {
RexNode compensationColumnsEquiPred = compensationPreds.getLeft();
RexNode otherCompensationPred = RexUtil.composeConjunction(rexBuilder, ImmutableList.of(compensationPreds.getMiddle(), compensationPreds.getRight()), false);
// a. Compute final compensation predicate.
if (!compensationColumnsEquiPred.isAlwaysTrue() || !otherCompensationPred.isAlwaysTrue()) {
// All columns required by compensating predicates must be contained
// in the view output (condition 2).
List<RexNode> viewExprs = topViewProject == null ? extractReferences(rexBuilder, view) : topViewProject.getChildExps();
// since we want to enforce the rest
if (!compensationColumnsEquiPred.isAlwaysTrue()) {
compensationColumnsEquiPred = rewriteExpression(rexBuilder, mq, view, viewNode, viewExprs, queryToViewTableMapping.inverse(), queryBasedVEC, false, compensationColumnsEquiPred);
if (compensationColumnsEquiPred == null) {
// Skip it
continue;
}
}
// For the rest, we use the query equivalence classes
if (!otherCompensationPred.isAlwaysTrue()) {
otherCompensationPred = rewriteExpression(rexBuilder, mq, view, viewNode, viewExprs, queryToViewTableMapping.inverse(), currQEC, true, otherCompensationPred);
if (otherCompensationPred == null) {
// Skip it
continue;
}
}
}
final RexNode viewCompensationPred = RexUtil.composeConjunction(rexBuilder, ImmutableList.of(compensationColumnsEquiPred, otherCompensationPred), false);
// b. Generate final rewriting if possible.
// First, we add the compensation predicate (if any) on top of the view.
// Then, we trigger the unifying method. This method will either create a
// Project or an Aggregate operator on top of the view. It will also compute
// the output expressions for the query.
RelBuilder builder = call.builder();
RelNode viewWithFilter;
if (!viewCompensationPred.isAlwaysTrue()) {
RexNode newPred = simplify.simplify(viewCompensationPred);
viewWithFilter = builder.push(view).filter(newPred).build();
// We add (and push) the filter to the view plan before triggering the rewriting.
// This is useful in case some of the columns can be folded to same value after
// filter is added.
Pair<RelNode, RelNode> pushedNodes = pushFilterToOriginalViewPlan(builder, topViewProject, viewNode, newPred);
topViewProject = (Project) pushedNodes.left;
viewNode = pushedNodes.right;
} else {
viewWithFilter = builder.push(view).build();
}
final RelNode result = rewriteView(builder, rexBuilder, simplify, mq, matchModality, false, viewWithFilter, topProject, node, topViewProject, viewNode, queryToViewTableMapping, currQEC);
if (result == null) {
// Skip it
continue;
}
call.transformTo(result);
}
// end else
}
}
}
}
}
use of org.apache.calcite.plan.RelOptMaterialization in project calcite by apache.
the class VolcanoPlanner method registerMaterializations.
private void registerMaterializations() {
// Avoid using materializations while populating materializations!
final CalciteConnectionConfig config = context.unwrap(CalciteConnectionConfig.class);
if (config == null || !config.materializationsEnabled()) {
return;
}
// Register rels using materialized views.
final List<Pair<RelNode, List<RelOptMaterialization>>> materializationUses = RelOptMaterializations.useMaterializedViews(originalRoot, materializations);
for (Pair<RelNode, List<RelOptMaterialization>> use : materializationUses) {
RelNode rel = use.left;
Hook.SUB.run(rel);
registerImpl(rel, root.set);
}
// Register table rels of materialized views that cannot find a substitution
// in root rel transformation but can potentially be useful.
final Set<RelOptMaterialization> applicableMaterializations = new HashSet<>(RelOptMaterializations.getApplicableMaterializations(originalRoot, materializations));
for (Pair<RelNode, List<RelOptMaterialization>> use : materializationUses) {
applicableMaterializations.removeAll(use.right);
}
for (RelOptMaterialization materialization : applicableMaterializations) {
RelSubset subset = registerImpl(materialization.queryRel, null);
RelNode tableRel2 = RelOptUtil.createCastRel(materialization.tableRel, materialization.queryRel.getRowType(), true);
registerImpl(tableRel2, subset.set);
}
// Register rels using lattices.
final List<Pair<RelNode, RelOptLattice>> latticeUses = RelOptMaterializations.useLattices(originalRoot, ImmutableList.copyOf(latticeByName.values()));
if (!latticeUses.isEmpty()) {
RelNode rel = latticeUses.get(0).left;
Hook.SUB.run(rel);
registerImpl(rel, root.set);
}
}
use of org.apache.calcite.plan.RelOptMaterialization in project hive by apache.
the class HiveMaterializedViewsRegistry method addMaterializedView.
/**
* Adds the materialized view to the cache.
*
* @param materializedViewTable the materialized view
*/
public RelOptMaterialization addMaterializedView(Table materializedViewTable) {
// Bail out if it is not enabled for rewriting
if (!materializedViewTable.isRewriteEnabled()) {
return null;
}
ConcurrentMap<ViewKey, RelOptMaterialization> cq = new ConcurrentHashMap<ViewKey, RelOptMaterialization>();
final ConcurrentMap<ViewKey, RelOptMaterialization> prevCq = materializedViews.putIfAbsent(materializedViewTable.getDbName(), cq);
if (prevCq != null) {
cq = prevCq;
}
// Bail out if it already exists
final ViewKey vk = new ViewKey(materializedViewTable.getTableName(), materializedViewTable.getCreateTime());
if (cq.containsKey(vk)) {
return null;
}
// Add to cache
final String viewQuery = materializedViewTable.getViewOriginalText();
final RelNode tableRel = createTableScan(materializedViewTable);
if (tableRel == null) {
LOG.warn("Materialized view " + materializedViewTable.getCompleteName() + " ignored; error creating view replacement");
return null;
}
final RelNode queryRel = parseQuery(viewQuery);
if (queryRel == null) {
LOG.warn("Materialized view " + materializedViewTable.getCompleteName() + " ignored; error parsing original query");
return null;
}
RelOptMaterialization materialization = new RelOptMaterialization(tableRel, queryRel, null);
cq.put(vk, materialization);
if (LOG.isDebugEnabled()) {
LOG.debug("Cached materialized view for rewriting: " + tableRel.getTable().getQualifiedName());
}
return materialization;
}
Aggregations