//////////////////////////////////////////////////////////////////////////////// /// DISCLAIMER /// /// Copyright 2014-2016 ArangoDB GmbH, Cologne, Germany /// Copyright 2004-2014 triAGENS GmbH, Cologne, Germany /// /// Licensed 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. /// /// Copyright holder is ArangoDB GmbH, Cologne, Germany /// /// @author Dr. Frank Celler //////////////////////////////////////////////////////////////////////////////// #include "MMFilesEdgeIndex.h" #include "Aql/AstNode.h" #include "Aql/SortCondition.h" #include "Basics/Exceptions.h" #include "Basics/LocalTaskQueue.h" #include "Basics/StaticStrings.h" #include "Basics/StringRef.h" #include "Basics/fasthash.h" #include "Basics/hashes.h" #include "Indexes/IndexLookupContext.h" #include "Indexes/SimpleAttributeEqualityMatcher.h" #include "MMFiles/MMFilesToken.h" #include "StorageEngine/TransactionState.h" #include "Utils/CollectionNameResolver.h" #include "Utils/TransactionMethods.h" #include "Utils/TransactionContext.h" #include "VocBase/LogicalCollection.h" #include #include using namespace arangodb; /// @brief hard-coded vector of the index attributes /// note that the attribute names must be hard-coded here to avoid an init-order /// fiasco with StaticStrings::FromString etc. static std::vector> const IndexAttributes{{arangodb::basics::AttributeName("_from", false)}, {arangodb::basics::AttributeName("_to", false)}}; /// @brief hashes an edge key static uint64_t HashElementKey(void*, VPackSlice const* key) { TRI_ASSERT(key != nullptr); // we can get away with the fast hash function here, as edge // index values are restricted to strings return MMFilesSimpleIndexElement::hash(*key); } /// @brief hashes an edge static uint64_t HashElementEdge(void*, MMFilesSimpleIndexElement const& element, bool byKey) { if (byKey) { return element.hash(); } TRI_voc_rid_t revisionId = element.revisionId(); return fasthash64_uint64(revisionId, 0x56781234); } /// @brief checks if key and element match static bool IsEqualKeyEdge(void* userData, VPackSlice const* left, MMFilesSimpleIndexElement const& right) { TRI_ASSERT(left != nullptr); IndexLookupContext* context = static_cast(userData); TRI_ASSERT(context != nullptr); try { VPackSlice tmp = right.slice(context); TRI_ASSERT(tmp.isString()); return left->equals(tmp); } catch (...) { return false; } } /// @brief checks for elements are equal static bool IsEqualElementEdge(void*, MMFilesSimpleIndexElement const& left, MMFilesSimpleIndexElement const& right) { return left.revisionId() == right.revisionId(); } /// @brief checks for elements are equal static bool IsEqualElementEdgeByKey(void* userData, MMFilesSimpleIndexElement const& left, MMFilesSimpleIndexElement const& right) { IndexLookupContext* context = static_cast(userData); try { VPackSlice lSlice = left.slice(context); VPackSlice rSlice = right.slice(context); TRI_ASSERT(lSlice.isString()); TRI_ASSERT(rSlice.isString()); return lSlice.equals(rSlice); } catch (...) { return false; } } MMFilesEdgeIndexIterator::MMFilesEdgeIndexIterator(LogicalCollection* collection, TransactionMethods* trx, ManagedDocumentResult* mmdr, arangodb::MMFilesEdgeIndex const* index, TRI_MMFilesEdgeIndexHash_t const* indexImpl, std::unique_ptr& keys) : IndexIterator(collection, trx, mmdr, index), _index(indexImpl), _keys(keys.get()), _iterator(_keys->slice()), _posInBuffer(0), _batchSize(1000), _lastElement() { keys.release(); // now we have ownership for _keys } MMFilesEdgeIndexIterator::~MMFilesEdgeIndexIterator() { if (_keys != nullptr) { // return the VPackBuilder to the transaction context _trx->transactionContextPtr()->returnBuilder(_keys.release()); } } bool MMFilesEdgeIndexIterator::next(TokenCallback const& cb, size_t limit) { if (limit == 0 || (_buffer.empty() && !_iterator.valid())) { // No limit no data, or we are actually done. The last call should have returned false TRI_ASSERT(limit > 0); // Someone called with limit == 0. Api broken return false; } while (limit > 0) { if (_buffer.empty()) { // We start a new lookup _posInBuffer = 0; VPackSlice tmp = _iterator.value(); if (tmp.isObject()) { tmp = tmp.get(StaticStrings::IndexEq); } _index->lookupByKey(&_context, &tmp, _buffer, _batchSize); } else if (_posInBuffer >= _buffer.size()) { // We have to refill the buffer _buffer.clear(); _posInBuffer = 0; _index->lookupByKeyContinue(&_context, _lastElement, _buffer, _batchSize); } if (_buffer.empty()) { _iterator.next(); _lastElement = MMFilesSimpleIndexElement(); if (!_iterator.valid()) { return false; } } else { _lastElement = _buffer.back(); // found something cb(MMFilesToken{_buffer[_posInBuffer++].revisionId()}); limit--; } } return true; } void MMFilesEdgeIndexIterator::reset() { _posInBuffer = 0; _buffer.clear(); _iterator.reset(); _lastElement = MMFilesSimpleIndexElement(); } MMFilesEdgeIndex::MMFilesEdgeIndex(TRI_idx_iid_t iid, arangodb::LogicalCollection* collection) : Index(iid, collection, std::vector>( {{arangodb::basics::AttributeName(StaticStrings::FromString, false)}, {arangodb::basics::AttributeName(StaticStrings::ToString, false)}}), false, false), _edgesFrom(nullptr), _edgesTo(nullptr), _numBuckets(1) { TRI_ASSERT(iid != 0); if (collection != nullptr) { // document is a nullptr in the coordinator case _numBuckets = static_cast(collection->indexBuckets()); } auto context = [this]() -> std::string { return this->context(); }; _edgesFrom = new TRI_MMFilesEdgeIndexHash_t(HashElementKey, HashElementEdge, IsEqualKeyEdge, IsEqualElementEdge, IsEqualElementEdgeByKey, _numBuckets, 64, context); _edgesTo = new TRI_MMFilesEdgeIndexHash_t( HashElementKey, HashElementEdge, IsEqualKeyEdge, IsEqualElementEdge, IsEqualElementEdgeByKey, _numBuckets, 64, context); } MMFilesEdgeIndex::~MMFilesEdgeIndex() { delete _edgesFrom; delete _edgesTo; } void MMFilesEdgeIndex::buildSearchValue(TRI_edge_direction_e dir, std::string const& id, VPackBuilder& builder) { builder.openArray(); switch (dir) { case TRI_EDGE_OUT: builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, VPackValue(id)); builder.close(); builder.close(); builder.add(VPackValue(VPackValueType::Null)); break; case TRI_EDGE_IN: builder.add(VPackValue(VPackValueType::Null)); builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, VPackValue(id)); builder.close(); builder.close(); break; case TRI_EDGE_ANY: builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, VPackValue(id)); builder.close(); builder.close(); builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, VPackValue(id)); builder.close(); builder.close(); } builder.close(); } void MMFilesEdgeIndex::buildSearchValue(TRI_edge_direction_e dir, VPackSlice const& id, VPackBuilder& builder) { TRI_ASSERT(id.isString()); builder.openArray(); switch (dir) { case TRI_EDGE_OUT: builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); builder.close(); builder.add(VPackValue(VPackValueType::Null)); break; case TRI_EDGE_IN: builder.add(VPackValue(VPackValueType::Null)); builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); builder.close(); break; case TRI_EDGE_ANY: builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); builder.close(); builder.openArray(); builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); builder.close(); } builder.close(); } void MMFilesEdgeIndex::buildSearchValueFromArray(TRI_edge_direction_e dir, VPackSlice const ids, VPackBuilder& builder) { TRI_ASSERT(ids.isArray()); builder.openArray(); switch (dir) { case TRI_EDGE_OUT: builder.openArray(); for (auto const& id : VPackArrayIterator(ids)) { if (id.isString()) { builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); } } builder.close(); builder.add(VPackValue(VPackValueType::Null)); break; case TRI_EDGE_IN: builder.add(VPackValue(VPackValueType::Null)); builder.openArray(); for (auto const& id : VPackArrayIterator(ids)) { if (id.isString()) { builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); } } builder.close(); break; case TRI_EDGE_ANY: builder.openArray(); for (auto const& id : VPackArrayIterator(ids)) { if (id.isString()) { builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); } } builder.close(); builder.openArray(); for (auto const& id : VPackArrayIterator(ids)) { if (id.isString()) { builder.openObject(); builder.add(StaticStrings::IndexEq, id); builder.close(); } } builder.close(); } builder.close(); } /// @brief return a selectivity estimate for the index double MMFilesEdgeIndex::selectivityEstimate(arangodb::StringRef const* attribute) const { if (_edgesFrom == nullptr || _edgesTo == nullptr || ServerState::instance()->isCoordinator()) { // use hard-coded selectivity estimate in case of cluster coordinator return 0.1; } if (attribute != nullptr) { // the index attribute is given here // now check if we can restrict the selectivity estimation to the correct // part of the index if (attribute->compare(StaticStrings::FromString) == 0) { // _from return _edgesFrom->selectivity(); } else if (attribute->compare(StaticStrings::ToString) == 0) { // _to return _edgesTo->selectivity(); } // other attribute. now return the average selectivity } // return average selectivity of the two index parts double estimate = (_edgesFrom->selectivity() + _edgesTo->selectivity()) * 0.5; TRI_ASSERT(estimate >= 0.0 && estimate <= 1.00001); // floating-point tolerance return estimate; } /// @brief return the memory usage for the index size_t MMFilesEdgeIndex::memory() const { TRI_ASSERT(_edgesFrom != nullptr); TRI_ASSERT(_edgesTo != nullptr); return _edgesFrom->memoryUsage() + _edgesTo->memoryUsage(); } /// @brief return a VelocyPack representation of the index void MMFilesEdgeIndex::toVelocyPack(VPackBuilder& builder, bool withFigures) const { Index::toVelocyPack(builder, withFigures); // hard-coded builder.add("unique", VPackValue(false)); builder.add("sparse", VPackValue(false)); } /// @brief return a VelocyPack representation of the index figures void MMFilesEdgeIndex::toVelocyPackFigures(VPackBuilder& builder) const { Index::toVelocyPackFigures(builder); builder.add("from", VPackValue(VPackValueType::Object)); _edgesFrom->appendToVelocyPack(builder); builder.close(); builder.add("to", VPackValue(VPackValueType::Object)); _edgesTo->appendToVelocyPack(builder); builder.close(); // builder.add("buckets", VPackValue(_numBuckets)); } int MMFilesEdgeIndex::insert(TransactionMethods* trx, TRI_voc_rid_t revisionId, VPackSlice const& doc, bool isRollback) { MMFilesSimpleIndexElement fromElement(buildFromElement(revisionId, doc)); MMFilesSimpleIndexElement toElement(buildToElement(revisionId, doc)); ManagedDocumentResult result; IndexLookupContext context(trx, _collection, &result, 1); _edgesFrom->insert(&context, fromElement, true, isRollback); try { _edgesTo->insert(&context, toElement, true, isRollback); } catch (...) { // roll back partial insert _edgesFrom->remove(&context, fromElement); return TRI_ERROR_OUT_OF_MEMORY; } return TRI_ERROR_NO_ERROR; } int MMFilesEdgeIndex::remove(TransactionMethods* trx, TRI_voc_rid_t revisionId, VPackSlice const& doc, bool isRollback) { MMFilesSimpleIndexElement fromElement(buildFromElement(revisionId, doc)); MMFilesSimpleIndexElement toElement(buildToElement(revisionId, doc)); ManagedDocumentResult result; IndexLookupContext context(trx, _collection, &result, 1); try { _edgesFrom->remove(&context, fromElement); _edgesTo->remove(&context, toElement); return TRI_ERROR_NO_ERROR; } catch (...) { if (isRollback) { return TRI_ERROR_NO_ERROR; } return TRI_ERROR_ARANGO_DOCUMENT_NOT_FOUND; } } void MMFilesEdgeIndex::batchInsert(TransactionMethods* trx, std::vector> const& documents, arangodb::basics::LocalTaskQueue* queue) { if (documents.empty()) { return; } std::shared_ptr> fromElements; fromElements.reset(new std::vector()); fromElements->reserve(documents.size()); std::shared_ptr> toElements; toElements.reset(new std::vector()); toElements->reserve(documents.size()); // functions that will be called for each thread auto creator = [&trx, this]() -> void* { ManagedDocumentResult* result = new ManagedDocumentResult; return new IndexLookupContext(trx, _collection, result, 1); }; auto destroyer = [](void* userData) { IndexLookupContext* context = static_cast(userData); delete context->result(); delete context; }; // TODO: create parallel tasks for this // _from for (auto const& it : documents) { VPackSlice value(TransactionMethods::extractFromFromDocument(it.second)); fromElements->emplace_back(MMFilesSimpleIndexElement( it.first, value, static_cast(value.begin() - it.second.begin()))); } // _to for (auto const& it : documents) { VPackSlice value(TransactionMethods::extractToFromDocument(it.second)); toElements->emplace_back(MMFilesSimpleIndexElement( it.first, value, static_cast(value.begin() - it.second.begin()))); } _edgesFrom->batchInsert(creator, destroyer, fromElements, queue); _edgesTo->batchInsert(creator, destroyer, toElements, queue); } /// @brief unload the index data from memory int MMFilesEdgeIndex::unload() { _edgesFrom->truncate([](MMFilesSimpleIndexElement const&) { return true; }); _edgesTo->truncate([](MMFilesSimpleIndexElement const&) { return true; }); return TRI_ERROR_NO_ERROR; } /// @brief provides a size hint for the edge index int MMFilesEdgeIndex::sizeHint(TransactionMethods* trx, size_t size) { // we assume this is called when setting up the index and the index // is still empty TRI_ASSERT(_edgesFrom->size() == 0); // set an initial size for the index for some new nodes to be created // without resizing ManagedDocumentResult result; IndexLookupContext context(trx, _collection, &result, 1); int err = _edgesFrom->resize(&context, size + 2049); if (err != TRI_ERROR_NO_ERROR) { return err; } // we assume this is called when setting up the index and the index // is still empty TRI_ASSERT(_edgesTo->size() == 0); // set an initial size for the index for some new nodes to be created // without resizing return _edgesTo->resize(&context, size + 2049); } /// @brief checks whether the index supports the condition bool MMFilesEdgeIndex::supportsFilterCondition( arangodb::aql::AstNode const* node, arangodb::aql::Variable const* reference, size_t itemsInIndex, size_t& estimatedItems, double& estimatedCost) const { SimpleAttributeEqualityMatcher matcher(IndexAttributes); return matcher.matchOne(this, node, reference, itemsInIndex, estimatedItems, estimatedCost); } /// @brief creates an IndexIterator for the given Condition IndexIterator* MMFilesEdgeIndex::iteratorForCondition( TransactionMethods* trx, ManagedDocumentResult* mmdr, arangodb::aql::AstNode const* node, arangodb::aql::Variable const* reference, bool reverse) const { TRI_ASSERT(node->type == aql::NODE_TYPE_OPERATOR_NARY_AND); TRI_ASSERT(node->numMembers() == 1); auto comp = node->getMember(0); // assume a.b == value auto attrNode = comp->getMember(0); auto valNode = comp->getMember(1); if (attrNode->type != aql::NODE_TYPE_ATTRIBUTE_ACCESS) { // got value == a.b -> flip sides attrNode = comp->getMember(1); valNode = comp->getMember(0); } TRI_ASSERT(attrNode->type == aql::NODE_TYPE_ATTRIBUTE_ACCESS); if (comp->type == aql::NODE_TYPE_OPERATOR_BINARY_EQ) { // a.b == value return createEqIterator(trx, mmdr, attrNode, valNode); } if (comp->type == aql::NODE_TYPE_OPERATOR_BINARY_IN) { // a.b IN values if (!valNode->isArray()) { return nullptr; } return createInIterator(trx, mmdr, attrNode, valNode); } // operator type unsupported return nullptr; } /// @brief specializes the condition for use with the index arangodb::aql::AstNode* MMFilesEdgeIndex::specializeCondition( arangodb::aql::AstNode* node, arangodb::aql::Variable const* reference) const { SimpleAttributeEqualityMatcher matcher(IndexAttributes); return matcher.specializeOne(this, node, reference); } /// @brief Transform the list of search slices to search values. /// This will multiply all IN entries and simply return all other /// entries. void MMFilesEdgeIndex::expandInSearchValues(VPackSlice const slice, VPackBuilder& builder) const { TRI_ASSERT(slice.isArray()); builder.openArray(); for (auto const& side : VPackArrayIterator(slice)) { if (side.isNull()) { builder.add(side); } else { TRI_ASSERT(side.isArray()); builder.openArray(); for (auto const& item : VPackArrayIterator(side)) { TRI_ASSERT(item.isObject()); if (item.hasKey(StaticStrings::IndexEq)) { TRI_ASSERT(!item.hasKey(StaticStrings::IndexIn)); builder.add(item); } else { TRI_ASSERT(item.hasKey(StaticStrings::IndexIn)); VPackSlice list = item.get(StaticStrings::IndexIn); TRI_ASSERT(list.isArray()); for (auto const& it : VPackArrayIterator(list)) { builder.openObject(); builder.add(StaticStrings::IndexEq, it); builder.close(); } } } builder.close(); } } builder.close(); } /// @brief create the iterator IndexIterator* MMFilesEdgeIndex::createEqIterator( TransactionMethods* trx, ManagedDocumentResult* mmdr, arangodb::aql::AstNode const* attrNode, arangodb::aql::AstNode const* valNode) const { // lease builder, but immediately pass it to the unique_ptr so we don't leak TransactionBuilderLeaser builder(trx); std::unique_ptr keys(builder.steal()); keys->openArray(); handleValNode(keys.get(), valNode); TRI_IF_FAILURE("EdgeIndex::noIterator") { THROW_ARANGO_EXCEPTION(TRI_ERROR_DEBUG); } keys->close(); // _from or _to? bool const isFrom = (attrNode->stringEquals(StaticStrings::FromString)); return new MMFilesEdgeIndexIterator(_collection, trx, mmdr, this, isFrom ? _edgesFrom : _edgesTo, keys); } /// @brief create the iterator IndexIterator* MMFilesEdgeIndex::createInIterator( TransactionMethods* trx, ManagedDocumentResult* mmdr, arangodb::aql::AstNode const* attrNode, arangodb::aql::AstNode const* valNode) const { // lease builder, but immediately pass it to the unique_ptr so we don't leak TransactionBuilderLeaser builder(trx); std::unique_ptr keys(builder.steal()); keys->openArray(); size_t const n = valNode->numMembers(); for (size_t i = 0; i < n; ++i) { handleValNode(keys.get(), valNode->getMemberUnchecked(i)); TRI_IF_FAILURE("EdgeIndex::iteratorValNodes") { THROW_ARANGO_EXCEPTION(TRI_ERROR_DEBUG); } } TRI_IF_FAILURE("EdgeIndex::noIterator") { THROW_ARANGO_EXCEPTION(TRI_ERROR_DEBUG); } keys->close(); // _from or _to? bool const isFrom = (attrNode->stringEquals(StaticStrings::FromString)); return new MMFilesEdgeIndexIterator(_collection, trx, mmdr, this, isFrom ? _edgesFrom : _edgesTo, keys); } /// @brief add a single value node to the iterator's keys void MMFilesEdgeIndex::handleValNode(VPackBuilder* keys, arangodb::aql::AstNode const* valNode) const { if (!valNode->isStringValue() || valNode->getStringLength() == 0) { return; } keys->openObject(); keys->add(StaticStrings::IndexEq, VPackValuePair(valNode->getStringValue(), valNode->getStringLength(), VPackValueType::String)); keys->close(); TRI_IF_FAILURE("EdgeIndex::collectKeys") { THROW_ARANGO_EXCEPTION(TRI_ERROR_DEBUG); } } MMFilesSimpleIndexElement MMFilesEdgeIndex::buildFromElement(TRI_voc_rid_t revisionId, VPackSlice const& doc) const { TRI_ASSERT(doc.isObject()); VPackSlice value(TransactionMethods::extractFromFromDocument(doc)); TRI_ASSERT(value.isString()); return MMFilesSimpleIndexElement(revisionId, value, static_cast(value.begin() - doc.begin())); } MMFilesSimpleIndexElement MMFilesEdgeIndex::buildToElement(TRI_voc_rid_t revisionId, VPackSlice const& doc) const { TRI_ASSERT(doc.isObject()); VPackSlice value(TransactionMethods::extractToFromDocument(doc)); TRI_ASSERT(value.isString()); return MMFilesSimpleIndexElement(revisionId, value, static_cast(value.begin() - doc.begin())); }