Class OverlapsJoinHashTable¶

class OverlapsJoinHashTable : public HashJoin ¶

Subclassed by RangeJoinHashTable

Public Functions

OverlapsJoinHashTable(const std::shared_ptr<Analyzer::BinOper> condition, const JoinType join_type, const std::vector<InputTableInfo> &query_infos, const Data_Namespace::MemoryLevel memory_level, ColumnCacheMap &column_cache, Executor *executor, const std::vector<InnerOuter> &inner_outer_pairs, const int device_count, HashtableAccessPathInfo hashtable_access_path_info, const TableIdToNodeMap &table_id_to_node_map)¶

virtual ~OverlapsJoinHashTable()¶

Public Static Functions

std::shared_ptr<OverlapsJoinHashTable> getInstance(const std::shared_ptr<Analyzer::BinOper> condition, const std::vector<InputTableInfo> &query_infos, const Data_Namespace::MemoryLevel memory_level, const JoinType join_type, const int device_count, ColumnCacheMap &column_cache, Executor *executor, const HashTableBuildDagMap &hashtable_build_dag_map, const RegisteredQueryHint &query_hint, const TableIdToNodeMap &table_id_to_node_map)¶: Make hash table from an in-flight SQL query’s parse tree etc.

static void invalidateCache()¶

static void markCachedItemAsDirty(size_t table_key)¶

static HashtableRecycler *getHashTableCache()¶

static OverlapsTuningParamRecycler *getOverlapsTuningParamCache()¶

Protected Functions

void reify(const HashType preferred_layout)¶

void reifyWithLayout(const HashType layout)¶

void reifyImpl(std::vector<ColumnsForDevice> &columns_per_device, const Fragmenter_Namespace::TableInfo &query_info, const HashType layout, const size_t shard_count, const size_t entry_count, const size_t emitted_keys_count, const bool skip_hashtable_caching, const size_t chosen_max_hashtable_size, const double chosen_bucket_threshold)¶

void reifyForDevice(const ColumnsForDevice &columns_for_device, const HashType layout, const size_t entry_count, const size_t emitted_keys_count, const bool skip_hashtable_caching, const int device_id, const logger::ThreadId parent_thread_id)¶

size_t calculateHashTableSize(size_t number_of_dimensions, size_t emitted_keys_count, size_t entry_count) const¶

ColumnsForDevice fetchColumnsForDevice(const std::vector<Fragmenter_Namespace::FragmentInfo> &fragments, const int device_id, DeviceAllocator *dev_buff_owner)¶

std::pair<size_t, size_t> approximateTupleCount(const std::vector<double> &inverse_bucket_sizes_for_dimension, std::vector<ColumnsForDevice> &columns_per_device, const size_t chosen_max_hashtable_size, const double chosen_bucket_threshold)¶

std::pair<size_t, size_t> computeHashTableCounts(const size_t shard_count, const std::vector<double> &inverse_bucket_sizes_for_dimension, std::vector<ColumnsForDevice> &columns_per_device, const size_t chosen_max_hashtable_size, const double chosen_bucket_threshold)¶

void setInverseBucketSizeInfo(const std::vector<double> &inverse_bucket_sizes, std::vector<ColumnsForDevice> &columns_per_device, const size_t device_count)¶

size_t getKeyComponentWidth() const¶

size_t getKeyComponentCount() const¶

HashType getHashType() const¶

Data_Namespace::MemoryLevel getMemoryLevel() const¶

int getDeviceCount() const¶

std::shared_ptr<BaselineHashTable> initHashTableOnCpu(const std::vector<JoinColumn> &join_columns, const std::vector<JoinColumnTypeInfo> &join_column_types, const std::vector<JoinBucketInfo> &join_bucket_info, const HashType layout, const size_t entry_count, const size_t emitted_keys_count, const bool skip_hashtable_caching)¶

HashJoinMatchingSet codegenMatchingSet(const CompilationOptions &co, const size_t index)¶

std::string toString(const ExecutorDeviceType device_type, const int device_id = 0, bool raw = false) const¶

std::set<DecodedJoinHashBufferEntry> toSet(const ExecutorDeviceType device_type, const int device_id) const¶

llvm::Value *codegenSlot(const CompilationOptions&, const size_t)¶

const RegisteredQueryHint &getRegisteredQueryHint()¶

void registerQueryHint(const RegisteredQueryHint &query_hint)¶

size_t getEntryCount() const¶

size_t getEmittedKeysCount() const¶

size_t getComponentBufferSize() const¶

size_t shardCount() const¶

Data_Namespace::MemoryLevel getEffectiveMemoryLevel(const std::vector<InnerOuter> &inner_outer_pairs) const¶

int getInnerTableId() const¶

int getInnerTableRteIdx() const¶

size_t getKeyBufferSize() const¶

size_t offsetBufferOff() const¶

size_t countBufferOff() const¶

size_t payloadBufferOff() const¶

std::string getHashJoinType() const¶

bool isBitwiseEq() const¶

std::shared_ptr<HashTable> initHashTableOnCpuFromCache(QueryPlanHash key, CacheItemType item_type, DeviceIdentifier device_identifier)¶

std::optional<std::pair<size_t, size_t>> getApproximateTupleCountFromCache(QueryPlanHash key, CacheItemType item_type, DeviceIdentifier device_identifier)¶

void putHashTableOnCpuToCache(QueryPlanHash key, CacheItemType item_type, std::shared_ptr<HashTable> hashtable_ptr, DeviceIdentifier device_identifier, size_t hashtable_building_time)¶

llvm::Value *codegenKey(const CompilationOptions &co)¶

std::vector<llvm::Value *> codegenManyKey(const CompilationOptions &co)¶

std::optional<OverlapsHashTableMetaInfo> getOverlapsHashTableMetaInfo()¶

QueryPlanHash getAlternativeCacheKey(AlternativeCacheKeyForOverlapsHashJoin &info)¶

void generateCacheKey(const size_t max_hashtable_size, const double bucket_threshold, const std::vector<double> &bucket_sizes)¶

QueryPlanHash getCacheKey() const¶

const std::vector<InnerOuter> &getInnerOuterPairs() const¶

void setOverlapsHashtableMetaInfo(size_t max_table_size_bytes, double bucket_threshold, std::vector<double> &bucket_sizes)¶

Protected Attributes

const std::shared_ptr<Analyzer::BinOper> condition_¶

const JoinType join_type_¶

const std::vector<InputTableInfo> &query_infos_¶

const Data_Namespace::MemoryLevel memory_level_¶

Executor *executor_¶

ColumnCacheMap &column_cache_¶

std::vector<InnerOuter> inner_outer_pairs_¶

const int device_count_¶

std::vector<double> inverse_bucket_sizes_for_dimension_¶

double chosen_overlaps_bucket_threshold_¶

size_t chosen_overlaps_max_table_size_bytes_¶

CompositeKeyInfo composite_key_info_¶

std::optional<HashType> layout_override_¶

std::mutex cpu_hash_table_buff_mutex_¶

RegisteredQueryHint query_hint_¶

QueryPlanDAG query_plan_dag_¶

QueryPlanHash hashtable_cache_key_¶

HashtableCacheMetaInfo hashtable_cache_meta_info_¶

std::unordered_set<size_t> table_keys_¶

const TableIdToNodeMap table_id_to_node_map_¶

Protected Static Attributes

std::unique_ptr<HashtableRecycler> hash_table_cache_ = std::make_unique<HashtableRecycler>(CacheItemType::OVERLAPS_HT, )¶

std::unique_ptr<OverlapsTuningParamRecycler> auto_tuner_cache_ = std::make_unique<OverlapsTuningParamRecycler>()¶

struct AlternativeCacheKeyForOverlapsHashJoin¶

Public Members

std::vector<InnerOuter> inner_outer_pairs¶

const size_t num_elements¶

const std::vector<ChunkKey> chunk_key¶

const SQLOps optype¶

const size_t max_hashtable_size¶

const double bucket_threshold¶

const std::vector<double> inverse_bucket_sizes = {}¶