duckdb 0.8.2-dev37.0 → 0.8.2-dev3989.0

package/src/duckdb/src/execution/aggregate_hashtable.cpp

@@ -7,13 +7,9 @@
 #include "duckdb/common/row_operations/row_operations.hpp"
 #include "duckdb/common/types/null_value.hpp"
 #include "duckdb/common/types/row/tuple_data_iterator.hpp"
-#include "duckdb/common/vector_operations/unary_executor.hpp"
 #include "duckdb/common/vector_operations/vector_operations.hpp"
 #include "duckdb/execution/expression_executor.hpp"
 #include "duckdb/planner/expression/bound_aggregate_expression.hpp"
-#include "duckdb/storage/buffer_manager.hpp"
-
-#include <cmath>
 
 namespace duckdb {
 
@@ -22,9 +18,9 @@ using ValidityBytes = TupleDataLayout::ValidityBytes;
 GroupedAggregateHashTable::GroupedAggregateHashTable(ClientContext &context, Allocator &allocator,
                                                      vector<LogicalType> group_types, vector<LogicalType> payload_types,
                                                      const vector<BoundAggregateExpression *> &bindings,
-                                                     HtEntryType entry_type, idx_t initial_capacity)
+                                                     idx_t initial_capacity, idx_t radix_bits)
     : GroupedAggregateHashTable(context, allocator, std::move(group_types), std::move(payload_types),
-                                AggregateObject::CreateAggregateObjects(bindings), entry_type, initial_capacity) {
+                                AggregateObject::CreateAggregateObjects(bindings), initial_capacity, radix_bits) {
 }
 
 GroupedAggregateHashTable::GroupedAggregateHashTable(ClientContext &context, Allocator &allocator,
@@ -32,51 +28,55 @@ GroupedAggregateHashTable::GroupedAggregateHashTable(ClientContext &context, All
     : GroupedAggregateHashTable(context, allocator, std::move(group_types), {}, vector<AggregateObject>()) {
 }
 
-AggregateHTAppendState::AggregateHTAppendState()
-    : ht_offsets(LogicalTypeId::BIGINT), hash_salts(LogicalTypeId::SMALLINT),
-      group_compare_vector(STANDARD_VECTOR_SIZE), no_match_vector(STANDARD_VECTOR_SIZE),
-      empty_vector(STANDARD_VECTOR_SIZE), new_groups(STANDARD_VECTOR_SIZE), addresses(LogicalType::POINTER),
-      chunk_state_initialized(false) {
+GroupedAggregateHashTable::AggregateHTAppendState::AggregateHTAppendState()
+    : ht_offsets(LogicalType::UBIGINT), hash_salts(LogicalType::HASH), group_compare_vector(STANDARD_VECTOR_SIZE),
+      no_match_vector(STANDARD_VECTOR_SIZE), empty_vector(STANDARD_VECTOR_SIZE), new_groups(STANDARD_VECTOR_SIZE),
+      addresses(LogicalType::POINTER) {
 }
 
 GroupedAggregateHashTable::GroupedAggregateHashTable(ClientContext &context, Allocator &allocator,
                                                      vector<LogicalType> group_types_p,
                                                      vector<LogicalType> payload_types_p,
                                                      vector<AggregateObject> aggregate_objects_p,
-                                                     HtEntryType entry_type, idx_t initial_capacity)
+                                                     idx_t initial_capacity, idx_t radix_bits)
     : BaseAggregateHashTable(context, allocator, aggregate_objects_p, std::move(payload_types_p)),
-      entry_type(entry_type), capacity(0), is_finalized(false),
-      aggregate_allocator(make_shared<ArenaAllocator>(allocator)) {
+      radix_bits(radix_bits), count(0), capacity(0), aggregate_allocator(make_shared<ArenaAllocator>(allocator)) {
+
 	// Append hash column to the end and initialise the row layout
 	group_types_p.emplace_back(LogicalType::HASH);
 	layout.Initialize(std::move(group_types_p), std::move(aggregate_objects_p));
-	tuple_size = layout.GetRowWidth();
-	tuples_per_block = Storage::BLOCK_SIZE / tuple_size;
-
-	// HT layout
 	hash_offset = layout.GetOffsets()[layout.ColumnCount() - 1];
-	data_collection = make_uniq<TupleDataCollection>(buffer_manager, layout);
-	data_collection->InitializeAppend(td_pin_state, TupleDataPinProperties::KEEP_EVERYTHING_PINNED);
 
-	hashes_hdl = buffer_manager.Allocate(Storage::BLOCK_SIZE);
-	hashes_hdl_ptr = hashes_hdl.Ptr();
+	// Partitioned data and pointer table
+	InitializePartitionedData();
+	Resize(initial_capacity);
 
-	switch (entry_type) {
-	case HtEntryType::HT_WIDTH_64: {
-		hash_prefix_shift = (HASH_WIDTH - sizeof(aggr_ht_entry_64::salt)) * 8;
-		Resize<aggr_ht_entry_64>(initial_capacity);
-		break;
-	}
-	case HtEntryType::HT_WIDTH_32: {
-		hash_prefix_shift = (HASH_WIDTH - sizeof(aggr_ht_entry_32::salt)) * 8;
-		Resize<aggr_ht_entry_32>(initial_capacity);
-		break;
-	}
-	default:
-		throw InternalException("Unknown HT entry width");
+	// Predicates
+	predicates.resize(layout.ColumnCount() - 1, ExpressionType::COMPARE_EQUAL);
+}
+
+void GroupedAggregateHashTable::InitializePartitionedData() {
+	if (!partitioned_data || RadixPartitioning::RadixBits(partitioned_data->PartitionCount()) != radix_bits) {
+		D_ASSERT(!partitioned_data || partitioned_data->Count() == 0);
+		partitioned_data =
+		    make_uniq<RadixPartitionedTupleData>(buffer_manager, layout, radix_bits, layout.ColumnCount() - 1);
+	} else {
+		partitioned_data->Reset();
 	}
 
-	predicates.resize(layout.ColumnCount() - 1, ExpressionType::COMPARE_EQUAL);
+	D_ASSERT(GetLayout().GetAggrWidth() == layout.GetAggrWidth());
+	D_ASSERT(GetLayout().GetDataWidth() == layout.GetDataWidth());
+	D_ASSERT(GetLayout().GetRowWidth() == layout.GetRowWidth());
+
+	partitioned_data->InitializeAppendState(state.append_state, TupleDataPinProperties::KEEP_EVERYTHING_PINNED);
+}
+
+unique_ptr<PartitionedTupleData> &GroupedAggregateHashTable::GetPartitionedData() {
+	return partitioned_data;
+}
+
+shared_ptr<ArenaAllocator> GroupedAggregateHashTable::GetAggregateAllocator() {
+	return aggregate_allocator;
 }
 
 GroupedAggregateHashTable::~GroupedAggregateHashTable() {
@@ -84,158 +84,133 @@ GroupedAggregateHashTable::~GroupedAggregateHashTable() {
 }
 
 void GroupedAggregateHashTable::Destroy() {
-	if (data_collection->Count() == 0) {
-		return;
-	}
-
-	// Check if there is an aggregate with a destructor
-	bool has_destructor = false;
-	for (auto &aggr : layout.GetAggregates()) {
-		if (aggr.function.destructor) {
-			has_destructor = true;
-		}
-	}
-	if (!has_destructor) {
+	if (!partitioned_data || partitioned_data->Count() == 0 || !layout.HasDestructor()) {
 		return;
 	}
 
 	// There are aggregates with destructors: Call the destructor for each of the aggregates
-	RowOperationsState state(aggregate_allocator->GetAllocator());
-	TupleDataChunkIterator iterator(*data_collection, TupleDataPinProperties::DESTROY_AFTER_DONE, false);
-	auto &row_locations = iterator.GetChunkState().row_locations;
-	do {
-		RowOperations::DestroyStates(state, layout, row_locations, iterator.GetCurrentChunkCount());
-	} while (iterator.Next());
-	data_collection->Reset();
-}
-
-template <class ENTRY>
-void GroupedAggregateHashTable::VerifyInternal() {
-	auto hashes_ptr = (ENTRY *)hashes_hdl_ptr;
-	idx_t count = 0;
-	for (idx_t i = 0; i < capacity; i++) {
-		if (hashes_ptr[i].page_nr > 0) {
-			D_ASSERT(hashes_ptr[i].page_offset < tuples_per_block);
-			D_ASSERT(hashes_ptr[i].page_nr <= payload_hds_ptrs.size());
-			auto ptr = payload_hds_ptrs[hashes_ptr[i].page_nr - 1] + ((hashes_ptr[i].page_offset) * tuple_size);
-			auto hash = Load<hash_t>(ptr + hash_offset);
-			D_ASSERT((hashes_ptr[i].salt) == (hash >> hash_prefix_shift));
-
-			count++;
+	// Currently does not happen because aggregate destructors are called while scanning in RadixPartitionedHashTable
+	// LCOV_EXCL_START
+	RowOperationsState row_state(*aggregate_allocator);
+	for (auto &data_collection : partitioned_data->GetPartitions()) {
+		if (data_collection->Count() == 0) {
+			continue;
 		}
+		TupleDataChunkIterator iterator(*data_collection, TupleDataPinProperties::DESTROY_AFTER_DONE, false);
+		auto &row_locations = iterator.GetChunkState().row_locations;
+		do {
+			RowOperations::DestroyStates(row_state, layout, row_locations, iterator.GetCurrentChunkCount());
+		} while (iterator.Next());
+		data_collection->Reset();
 	}
-	(void)count;
-	D_ASSERT(count == Count());
+	// LCOV_EXCL_STOP
+}
+
+const TupleDataLayout &GroupedAggregateHashTable::GetLayout() const {
+	return partitioned_data->GetLayout();
+}
+
+idx_t GroupedAggregateHashTable::Count() const {
+	return count;
 }
 
 idx_t GroupedAggregateHashTable::InitialCapacity() {
 	return STANDARD_VECTOR_SIZE * 2ULL;
 }
 
-idx_t GroupedAggregateHashTable::GetMaxCapacity(HtEntryType entry_type, idx_t tuple_size) {
-	idx_t max_pages;
-	idx_t max_tuples;
-
-	switch (entry_type) {
-	case HtEntryType::HT_WIDTH_32:
-		max_pages = NumericLimits<uint8_t>::Maximum();
-		max_tuples = NumericLimits<uint16_t>::Maximum();
-		break;
-	case HtEntryType::HT_WIDTH_64:
-		max_pages = NumericLimits<uint32_t>::Maximum();
-		max_tuples = NumericLimits<uint16_t>::Maximum();
-		break;
-	default:
-		throw InternalException("Unsupported hash table width");
-	}
+idx_t GroupedAggregateHashTable::GetCapacityForCount(idx_t count) {
+	count = MaxValue<idx_t>(InitialCapacity(), count);
+	return NextPowerOfTwo(count * LOAD_FACTOR);
+}
+
+idx_t GroupedAggregateHashTable::Capacity() const {
+	return capacity;
+}
 
-	return max_pages * MinValue(max_tuples, (idx_t)Storage::BLOCK_SIZE / tuple_size);
+idx_t GroupedAggregateHashTable::ResizeThreshold() const {
+	return Capacity() / LOAD_FACTOR;
 }
 
-idx_t GroupedAggregateHashTable::MaxCapacity() {
-	return GetMaxCapacity(entry_type, tuple_size);
+idx_t GroupedAggregateHashTable::ApplyBitMask(hash_t hash) const {
+	return hash & bitmask;
 }
 
 void GroupedAggregateHashTable::Verify() {
 #ifdef DEBUG
-	switch (entry_type) {
-	case HtEntryType::HT_WIDTH_32:
-		VerifyInternal<aggr_ht_entry_32>();
-		break;
-	case HtEntryType::HT_WIDTH_64:
-		VerifyInternal<aggr_ht_entry_64>();
-		break;
+	idx_t total_count = 0;
+	for (idx_t i = 0; i < capacity; i++) {
+		const auto &entry = entries[i];
+		if (!entry.IsOccupied()) {
+			continue;
+		}
+		auto hash = Load<hash_t>(entry.GetPointer() + hash_offset);
+		D_ASSERT(entry.GetSalt() == aggr_ht_entry_t::ExtractSalt(hash));
+		total_count++;
 	}
+	D_ASSERT(total_count == Count());
 #endif
 }
 
-template <class ENTRY>
+void GroupedAggregateHashTable::ClearPointerTable() {
+	std::fill_n(entries, capacity, aggr_ht_entry_t(0));
+}
+
+void GroupedAggregateHashTable::ResetCount() {
+	count = 0;
+}
+
+void GroupedAggregateHashTable::SetRadixBits(idx_t radix_bits_p) {
+	radix_bits = radix_bits_p;
+}
+
 void GroupedAggregateHashTable::Resize(idx_t size) {
-	D_ASSERT(!is_finalized);
 	D_ASSERT(size >= STANDARD_VECTOR_SIZE);
 	D_ASSERT(IsPowerOfTwo(size));
-
 	if (size < capacity) {
 		throw InternalException("Cannot downsize a hash table!");
 	}
-	capacity = size;
 
+	capacity = size;
+	hash_map = buffer_manager.GetBufferAllocator().Allocate(capacity * sizeof(aggr_ht_entry_t));
+	entries = reinterpret_cast<aggr_ht_entry_t *>(hash_map.get());
+	ClearPointerTable();
 	bitmask = capacity - 1;
-	const auto byte_size = capacity * sizeof(ENTRY);
-	if (byte_size > (idx_t)Storage::BLOCK_SIZE) {
-		hashes_hdl = buffer_manager.Allocate(byte_size);
-		hashes_hdl_ptr = hashes_hdl.Ptr();
-	}
-	memset(hashes_hdl_ptr, 0, byte_size);
 
 	if (Count() != 0) {
-		D_ASSERT(!payload_hds_ptrs.empty());
-		auto hashes_arr = (ENTRY *)hashes_hdl_ptr;
-
-		idx_t block_id = 0;
-		auto block_pointer = payload_hds_ptrs[block_id];
-		auto block_end = block_pointer + tuples_per_block * tuple_size;
-
-		TupleDataChunkIterator iterator(*data_collection, TupleDataPinProperties::ALREADY_PINNED, false);
-		const auto row_locations = iterator.GetRowLocations();
-		do {
-			for (idx_t i = 0; i < iterator.GetCurrentChunkCount(); i++) {
-				const auto &row_location = row_locations[i];
-				if (row_location > block_end || row_location < block_pointer) {
-					block_id++;
-					D_ASSERT(block_id < payload_hds_ptrs.size());
-					block_pointer = payload_hds_ptrs[block_id];
-					block_end = block_pointer + tuples_per_block * tuple_size;
-				}
-				D_ASSERT(row_location >= block_pointer && row_location < block_end);
-				D_ASSERT((row_location - block_pointer) % tuple_size == 0);
-
-				const auto hash = Load<hash_t>(row_location + hash_offset);
-				D_ASSERT((hash & bitmask) == (hash % capacity));
-				D_ASSERT(hash >> hash_prefix_shift <= NumericLimits<uint16_t>::Maximum());
-
-				auto entry_idx = (idx_t)hash & bitmask;
-				while (hashes_arr[entry_idx].page_nr > 0) {
-					entry_idx++;
-					if (entry_idx >= capacity) {
-						entry_idx = 0;
+		for (auto &data_collection : partitioned_data->GetPartitions()) {
+			if (data_collection->Count() == 0) {
+				continue;
+			}
+			TupleDataChunkIterator iterator(*data_collection, TupleDataPinProperties::ALREADY_PINNED, false);
+			const auto row_locations = iterator.GetRowLocations();
+			do {
+				for (idx_t i = 0; i < iterator.GetCurrentChunkCount(); i++) {
+					const auto &row_location = row_locations[i];
+					const auto hash = Load<hash_t>(row_location + hash_offset);
+
+					// Find an empty entry
+					auto entry_idx = ApplyBitMask(hash);
+					D_ASSERT(entry_idx == hash % capacity);
+					while (entries[entry_idx].IsOccupied() > 0) {
+						entry_idx++;
+						if (entry_idx >= capacity) {
+							entry_idx = 0;
+						}
 					}
+					auto &entry = entries[entry_idx];
+					D_ASSERT(!entry.IsOccupied());
+					entry.SetSalt(aggr_ht_entry_t::ExtractSalt(hash));
+					entry.SetPointer(row_location);
+					D_ASSERT(entry.IsOccupied());
 				}
-
-				auto &ht_entry = hashes_arr[entry_idx];
-				D_ASSERT(!ht_entry.page_nr);
-				ht_entry.salt = hash >> hash_prefix_shift;
-				ht_entry.page_nr = block_id + 1;
-				ht_entry.page_offset = (row_location - block_pointer) / tuple_size;
-			}
-		} while (iterator.Next());
+			} while (iterator.Next());
+		}
 	}
 
 	Verify();
 }
 
-idx_t GroupedAggregateHashTable::AddChunk(AggregateHTAppendState &state, DataChunk &groups, DataChunk &payload,
-                                          AggregateType filter) {
+idx_t GroupedAggregateHashTable::AddChunk(DataChunk &groups, DataChunk &payload, AggregateType filter) {
 	unsafe_vector<idx_t> aggregate_filter;
 
 	auto &aggregates = layout.GetAggregates();
@@ -245,20 +220,18 @@ idx_t GroupedAggregateHashTable::AddChunk(AggregateHTAppendState &state, DataChu
 			aggregate_filter.push_back(i);
 		}
 	}
-	return AddChunk(state, groups, payload, aggregate_filter);
+	return AddChunk(groups, payload, aggregate_filter);
 }
 
-idx_t GroupedAggregateHashTable::AddChunk(AggregateHTAppendState &state, DataChunk &groups, DataChunk &payload,
-                                          const unsafe_vector<idx_t> &filter) {
+idx_t GroupedAggregateHashTable::AddChunk(DataChunk &groups, DataChunk &payload, const unsafe_vector<idx_t> &filter) {
 	Vector hashes(LogicalType::HASH);
 	groups.Hash(hashes);
 
-	return AddChunk(state, groups, hashes, payload, filter);
+	return AddChunk(groups, hashes, payload, filter);
 }
 
-idx_t GroupedAggregateHashTable::AddChunk(AggregateHTAppendState &state, DataChunk &groups, Vector &group_hashes,
-                                          DataChunk &payload, const unsafe_vector<idx_t> &filter) {
-	D_ASSERT(!is_finalized);
+idx_t GroupedAggregateHashTable::AddChunk(DataChunk &groups, Vector &group_hashes, DataChunk &payload,
+                                          const unsafe_vector<idx_t> &filter) {
 	if (groups.size() == 0) {
 		return 0;
 	}
@@ -270,14 +243,14 @@ idx_t GroupedAggregateHashTable::AddChunk(AggregateHTAppendState &state, DataChu
 	}
 #endif
 
-	auto new_group_count = FindOrCreateGroups(state, groups, group_hashes, state.addresses, state.new_groups);
+	const auto new_group_count = FindOrCreateGroups(groups, group_hashes, state.addresses, state.new_groups);
 	VectorOperations::AddInPlace(state.addresses, layout.GetAggrOffset(), payload.size());
 
 	// Now every cell has an entry, update the aggregates
 	auto &aggregates = layout.GetAggregates();
 	idx_t filter_idx = 0;
 	idx_t payload_idx = 0;
-	RowOperationsState row_state(aggregate_allocator->GetAllocator());
+	RowOperationsState row_state(*aggregate_allocator);
 	for (idx_t i = 0; i < aggregates.size(); i++) {
 		auto &aggr = aggregates[i];
 		if (filter_idx >= filter.size() || i < filter[filter_idx]) {
@@ -306,11 +279,14 @@ idx_t GroupedAggregateHashTable::AddChunk(AggregateHTAppendState &state, DataChu
 }
 
 void GroupedAggregateHashTable::FetchAggregates(DataChunk &groups, DataChunk &result) {
+#ifdef DEBUG
 	groups.Verify();
 	D_ASSERT(groups.ColumnCount() + 1 == layout.ColumnCount());
 	for (idx_t i = 0; i < result.ColumnCount(); i++) {
 		D_ASSERT(result.data[i].GetType() == payload_types[i]);
 	}
+#endif
+
 	result.SetCardinality(groups);
 	if (groups.size() == 0) {
 		return;
@@ -318,57 +294,46 @@ void GroupedAggregateHashTable::FetchAggregates(DataChunk &groups, DataChunk &re
 
 	// find the groups associated with the addresses
 	// FIXME: this should not use the FindOrCreateGroups, creating them is unnecessary
-	AggregateHTAppendState append_state;
 	Vector addresses(LogicalType::POINTER);
-	FindOrCreateGroups(append_state, groups, addresses);
+	FindOrCreateGroups(groups, addresses);
 	// now fetch the aggregates
-	RowOperationsState row_state(aggregate_allocator->GetAllocator());
+	RowOperationsState row_state(*aggregate_allocator);
 	RowOperations::FinalizeStates(row_state, layout, addresses, result, 0);
 }
 
-idx_t GroupedAggregateHashTable::ResizeThreshold() {
-	return capacity / LOAD_FACTOR;
-}
-
-template <class ENTRY>
-idx_t GroupedAggregateHashTable::FindOrCreateGroupsInternal(AggregateHTAppendState &state, DataChunk &groups,
-                                                            Vector &group_hashes_v, Vector &addresses_v,
-                                                            SelectionVector &new_groups_out) {
-	D_ASSERT(!is_finalized);
+idx_t GroupedAggregateHashTable::FindOrCreateGroupsInternal(DataChunk &groups, Vector &group_hashes_v,
+                                                            Vector &addresses_v, SelectionVector &new_groups_out) {
 	D_ASSERT(groups.ColumnCount() + 1 == layout.ColumnCount());
 	D_ASSERT(group_hashes_v.GetType() == LogicalType::HASH);
 	D_ASSERT(state.ht_offsets.GetVectorType() == VectorType::FLAT_VECTOR);
-	D_ASSERT(state.ht_offsets.GetType() == LogicalType::BIGINT);
+	D_ASSERT(state.ht_offsets.GetType() == LogicalType::UBIGINT);
 	D_ASSERT(addresses_v.GetType() == LogicalType::POINTER);
-	D_ASSERT(state.hash_salts.GetType() == LogicalType::SMALLINT);
+	D_ASSERT(state.hash_salts.GetType() == LogicalType::HASH);
 
-	if (Count() + groups.size() > MaxCapacity()) {
-		throw InternalException("Hash table capacity reached");
-	}
-
-	// Resize at 50% capacity, also need to fit the entire vector
-	if (capacity - Count() <= groups.size() || Count() > ResizeThreshold()) {
+	// Need to fit the entire vector, and resize at threshold
+	if (Count() + groups.size() > capacity || Count() + groups.size() > ResizeThreshold()) {
 		Verify();
-		Resize<ENTRY>(capacity * 2);
+		Resize(capacity * 2);
 	}
 	D_ASSERT(capacity - Count() >= groups.size()); // we need to be able to fit at least one vector of data
 
 	group_hashes_v.Flatten(groups.size());
-	auto group_hashes = FlatVector::GetData<hash_t>(group_hashes_v);
+	auto hashes = FlatVector::GetData<hash_t>(group_hashes_v);
 
 	addresses_v.Flatten(groups.size());
 	auto addresses = FlatVector::GetData<data_ptr_t>(addresses_v);
 
 	// Compute the entry in the table based on the hash using a modulo,
 	// and precompute the hash salts for faster comparison below
-	auto ht_offsets_ptr = FlatVector::GetData<uint64_t>(state.ht_offsets);
-	auto hash_salts_ptr = FlatVector::GetData<uint16_t>(state.hash_salts);
+	auto ht_offsets = FlatVector::GetData<uint64_t>(state.ht_offsets);
+	auto hash_salts = FlatVector::GetData<hash_t>(state.hash_salts);
 	for (idx_t r = 0; r < groups.size(); r++) {
-		auto element = group_hashes[r];
-		D_ASSERT((element & bitmask) == (element % capacity));
-		ht_offsets_ptr[r] = element & bitmask;
-		hash_salts_ptr[r] = element >> hash_prefix_shift;
+		const auto &hash = hashes[r];
+		ht_offsets[r] = ApplyBitMask(hash);
+		D_ASSERT(ht_offsets[r] == hash % capacity);
+		hash_salts[r] = aggr_ht_entry_t::ExtractSalt(hash);
 	}
+
 	// we start out with all entries [0, 1, 2, ..., groups.size()]
 	const SelectionVector *sel_vector = FlatVector::IncrementalSelectionVector();
 
@@ -384,15 +349,12 @@ idx_t GroupedAggregateHashTable::FindOrCreateGroupsInternal(AggregateHTAppendSta
 	state.group_chunk.SetCardinality(groups);
 
 	// convert all vectors to unified format
-	if (!state.chunk_state_initialized) {
-		data_collection->InitializeAppend(state.chunk_state);
-		state.chunk_state_initialized = true;
-	}
-	TupleDataCollection::ToUnifiedFormat(state.chunk_state, state.group_chunk);
+	auto &chunk_state = state.append_state.chunk_state;
+	TupleDataCollection::ToUnifiedFormat(chunk_state, state.group_chunk);
 	if (!state.group_data) {
 		state.group_data = make_unsafe_uniq_array<UnifiedVectorFormat>(state.group_chunk.ColumnCount());
 	}
-	TupleDataCollection::GetVectorData(state.chunk_state, state.group_data.get());
+	TupleDataCollection::GetVectorData(chunk_state, state.group_data.get());
 
 	idx_t new_group_count = 0;
 	idx_t remaining_entries = groups.size();
@@ -403,57 +365,42 @@ idx_t GroupedAggregateHashTable::FindOrCreateGroupsInternal(AggregateHTAppendSta
 
 		// For each remaining entry, figure out whether or not it belongs to a full or empty group
 		for (idx_t i = 0; i < remaining_entries; i++) {
-			const idx_t index = sel_vector->get_index(i);
-			auto &ht_entry = *(((ENTRY *)this->hashes_hdl_ptr) + ht_offsets_ptr[index]);
-			if (ht_entry.page_nr == 0) { // Cell is unoccupied (we use page number 0 as a "unused marker")
-				D_ASSERT(group_hashes[index] >> hash_prefix_shift <= NumericLimits<uint16_t>::Maximum());
-				D_ASSERT(payload_hds_ptrs.size() < NumericLimits<uint32_t>::Maximum());
-
-				// Set page nr to 1 for now to mark it as occupied (will be corrected later) and set the salt
-				ht_entry.page_nr = 1;
-				ht_entry.salt = group_hashes[index] >> hash_prefix_shift;
-
-				// Update selection lists for outer loops
-				state.empty_vector.set_index(new_entry_count++, index);
-				new_groups_out.set_index(new_group_count++, index);
-			} else { // Cell is occupied: Compare salts
-				if (ht_entry.salt == hash_salts_ptr[index]) {
+			const auto index = sel_vector->get_index(i);
+			const auto &salt = hash_salts[index];
+			auto &entry = entries[ht_offsets[index]];
+			if (entry.IsOccupied()) { // Cell is occupied: Compare salts
+				if (entry.GetSalt() == salt) {
 					state.group_compare_vector.set_index(need_compare_count++, index);
 				} else {
 					state.no_match_vector.set_index(no_match_count++, index);
 				}
+			} else { // Cell is unoccupied
+				// Set salt (also marks as occupied)
+				entry.SetSalt(salt);
+
+				// Update selection lists for outer loops
+				state.empty_vector.set_index(new_entry_count++, index);
+				new_groups_out.set_index(new_group_count++, index);
 			}
 		}
 
 		if (new_entry_count != 0) {
 			// Append everything that belongs to an empty group
-			data_collection->AppendUnified(td_pin_state, state.chunk_state, state.group_chunk, state.empty_vector,
-			                               new_entry_count);
-			RowOperations::InitializeStates(layout, state.chunk_state.row_locations,
+			partitioned_data->AppendUnified(state.append_state, state.group_chunk, state.empty_vector, new_entry_count);
+			RowOperations::InitializeStates(layout, chunk_state.row_locations,
 			                                *FlatVector::IncrementalSelectionVector(), new_entry_count);
 
-			// Get the pointers to the (possibly) newly created blocks of the data collection
-			idx_t block_id = payload_hds_ptrs.empty() ? 0 : payload_hds_ptrs.size() - 1;
-			UpdateBlockPointers();
-			auto block_pointer = payload_hds_ptrs[block_id];
-			auto block_end = block_pointer + tuples_per_block * tuple_size;
-
-			// Set the page nrs/offsets in the 1st part of the HT now that the data has been appended
-			const auto row_locations = FlatVector::GetData<data_ptr_t>(state.chunk_state.row_locations);
+			// Set the entry pointers in the 1st part of the HT now that the data has been appended
+			const auto row_locations = FlatVector::GetData<data_ptr_t>(chunk_state.row_locations);
+			const auto &row_sel = state.append_state.reverse_partition_sel;
 			for (idx_t new_entry_idx = 0; new_entry_idx < new_entry_count; new_entry_idx++) {
-				const auto &row_location = row_locations[new_entry_idx];
-				if (row_location > block_end || row_location < block_pointer) {
-					block_id++;
-					D_ASSERT(block_id < payload_hds_ptrs.size());
-					block_pointer = payload_hds_ptrs[block_id];
-					block_end = block_pointer + tuples_per_block * tuple_size;
-				}
-				D_ASSERT(row_location >= block_pointer && row_location < block_end);
-				D_ASSERT((row_location - block_pointer) % tuple_size == 0);
 				const auto index = state.empty_vector.get_index(new_entry_idx);
-				auto &ht_entry = *(((ENTRY *)this->hashes_hdl_ptr) + ht_offsets_ptr[index]);
-				ht_entry.page_nr = block_id + 1;
-				ht_entry.page_offset = (row_location - block_pointer) / tuple_size;
+				const auto row_idx = row_sel.get_index(index);
+				const auto &row_location = row_locations[row_idx];
+
+				auto &entry = entries[ht_offsets[index]];
+
+				entry.SetPointer(row_location);
 				addresses[index] = row_location;
 			}
 		}
@@ -462,10 +409,8 @@ idx_t GroupedAggregateHashTable::FindOrCreateGroupsInternal(AggregateHTAppendSta
 			// Get the pointers to the rows that need to be compared
 			for (idx_t need_compare_idx = 0; need_compare_idx < need_compare_count; need_compare_idx++) {
 				const auto index = state.group_compare_vector.get_index(need_compare_idx);
-				const auto &ht_entry = *(((ENTRY *)this->hashes_hdl_ptr) + ht_offsets_ptr[index]);
-				auto page_ptr = payload_hds_ptrs[ht_entry.page_nr - 1];
-				auto page_offset = ht_entry.page_offset * tuple_size;
-				addresses[index] = page_ptr + page_offset;
+				const auto &entry = entries[ht_offsets[index]];
+				addresses[index] = entry.GetPointer();
 			}
 
 			// Perform group comparisons
@@ -477,55 +422,36 @@ idx_t GroupedAggregateHashTable::FindOrCreateGroupsInternal(AggregateHTAppendSta
 		// Linear probing: each of the entries that do not match move to the next entry in the HT
 		for (idx_t i = 0; i < no_match_count; i++) {
 			idx_t index = state.no_match_vector.get_index(i);
-			ht_offsets_ptr[index]++;
-			if (ht_offsets_ptr[index] >= capacity) {
-				ht_offsets_ptr[index] = 0;
+			ht_offsets[index]++;
+			if (ht_offsets[index] >= capacity) {
+				ht_offsets[index] = 0;
 			}
 		}
 		sel_vector = &state.no_match_vector;
 		remaining_entries = no_match_count;
 	}
 
+	count += new_group_count;
 	return new_group_count;
 }
 
-void GroupedAggregateHashTable::UpdateBlockPointers() {
-	for (const auto &id_and_handle : td_pin_state.row_handles) {
-		const auto &id = id_and_handle.first;
-		const auto &handle = id_and_handle.second;
-		if (payload_hds_ptrs.empty() || id > payload_hds_ptrs.size() - 1) {
-			payload_hds_ptrs.resize(id + 1);
-		}
-		payload_hds_ptrs[id] = handle.Ptr();
-	}
-}
-
 // this is to support distinct aggregations where we need to record whether we
 // have already seen a value for a group
-idx_t GroupedAggregateHashTable::FindOrCreateGroups(AggregateHTAppendState &state, DataChunk &groups,
-                                                    Vector &group_hashes, Vector &addresses_out,
+idx_t GroupedAggregateHashTable::FindOrCreateGroups(DataChunk &groups, Vector &group_hashes, Vector &addresses_out,
                                                     SelectionVector &new_groups_out) {
-	switch (entry_type) {
-	case HtEntryType::HT_WIDTH_64:
-		return FindOrCreateGroupsInternal<aggr_ht_entry_64>(state, groups, group_hashes, addresses_out, new_groups_out);
-	case HtEntryType::HT_WIDTH_32:
-		return FindOrCreateGroupsInternal<aggr_ht_entry_32>(state, groups, group_hashes, addresses_out, new_groups_out);
-	default:
-		throw InternalException("Unknown HT entry width");
-	}
+	return FindOrCreateGroupsInternal(groups, group_hashes, addresses_out, new_groups_out);
 }
 
-void GroupedAggregateHashTable::FindOrCreateGroups(AggregateHTAppendState &state, DataChunk &groups,
-                                                   Vector &addresses) {
+void GroupedAggregateHashTable::FindOrCreateGroups(DataChunk &groups, Vector &addresses) {
 	// create a dummy new_groups sel vector
-	FindOrCreateGroups(state, groups, addresses, state.new_groups);
+	FindOrCreateGroups(groups, addresses, state.new_groups);
 }
 
-idx_t GroupedAggregateHashTable::FindOrCreateGroups(AggregateHTAppendState &state, DataChunk &groups,
-                                                    Vector &addresses_out, SelectionVector &new_groups_out) {
+idx_t GroupedAggregateHashTable::FindOrCreateGroups(DataChunk &groups, Vector &addresses_out,
+                                                    SelectionVector &new_groups_out) {
 	Vector hashes(LogicalType::HASH);
 	groups.Hash(hashes);
-	return FindOrCreateGroups(state, groups, hashes, addresses_out, new_groups_out);
+	return FindOrCreateGroups(groups, hashes, addresses_out, new_groups_out);
 }
 
 struct FlushMoveState {
@@ -538,13 +464,21 @@ struct FlushMoveState {
 		for (idx_t col_idx = 0; col_idx < layout.ColumnCount() - 1; col_idx++) {
 			column_ids.emplace_back(col_idx);
 		}
-		// FIXME DESTROY_AFTER_DONE if we make it possible to pass a selection vector to RowOperations::DestroyStates?
-		collection.InitializeScan(scan_state, column_ids, TupleDataPinProperties::UNPIN_AFTER_DONE);
+		collection.InitializeScan(scan_state, column_ids, TupleDataPinProperties::DESTROY_AFTER_DONE);
 		collection.InitializeScanChunk(scan_state, groups);
 		hash_col_idx = layout.ColumnCount() - 1;
 	}
 
-	bool Scan();
+	bool Scan() {
+		if (collection.Scan(scan_state, groups)) {
+			collection.Gather(scan_state.chunk_state.row_locations, *FlatVector::IncrementalSelectionVector(),
+			                  groups.size(), hash_col_idx, hashes, *FlatVector::IncrementalSelectionVector());
+			return true;
+		}
+
+		collection.FinalizePinState(scan_state.pin_state);
+		return false;
+	}
 
 	TupleDataCollection &collection;
 	TupleDataScanState scan_state;
@@ -553,102 +487,48 @@ struct FlushMoveState {
 	idx_t hash_col_idx;
 	Vector hashes;
 
-	AggregateHTAppendState append_state;
 	Vector group_addresses;
 	SelectionVector new_groups_sel;
 };
 
-bool FlushMoveState::Scan() {
-	if (collection.Scan(scan_state, groups)) {
-		collection.Gather(scan_state.chunk_state.row_locations, *FlatVector::IncrementalSelectionVector(),
-		                  groups.size(), hash_col_idx, hashes, *FlatVector::IncrementalSelectionVector());
-		return true;
-	}
+void GroupedAggregateHashTable::Combine(GroupedAggregateHashTable &other) {
+	auto other_data = other.partitioned_data->GetUnpartitioned();
+	Combine(*other_data);
 
-	collection.FinalizePinState(scan_state.pin_state);
-	return false;
+	// Inherit ownership to all stored aggregate allocators
+	stored_allocators.emplace_back(other.aggregate_allocator);
+	for (const auto &stored_allocator : other.stored_allocators) {
+		stored_allocators.emplace_back(stored_allocator);
+	}
 }
 
-void GroupedAggregateHashTable::Combine(GroupedAggregateHashTable &other) {
-	D_ASSERT(!is_finalized);
-
-	D_ASSERT(other.layout.GetAggrWidth() == layout.GetAggrWidth());
-	D_ASSERT(other.layout.GetDataWidth() == layout.GetDataWidth());
-	D_ASSERT(other.layout.GetRowWidth() == layout.GetRowWidth());
+void GroupedAggregateHashTable::Combine(TupleDataCollection &other_data) {
+	D_ASSERT(other_data.GetLayout().GetAggrWidth() == layout.GetAggrWidth());
+	D_ASSERT(other_data.GetLayout().GetDataWidth() == layout.GetDataWidth());
+	D_ASSERT(other_data.GetLayout().GetRowWidth() == layout.GetRowWidth());
 
-	if (other.Count() == 0) {
+	if (other_data.Count() == 0) {
 		return;
 	}
 
-	FlushMoveState state(*other.data_collection);
-	RowOperationsState row_state(aggregate_allocator->GetAllocator());
-	while (state.Scan()) {
-		FindOrCreateGroups(state.append_state, state.groups, state.hashes, state.group_addresses, state.new_groups_sel);
-		RowOperations::CombineStates(row_state, layout, state.scan_state.chunk_state.row_locations,
-		                             state.group_addresses, state.groups.size());
+	FlushMoveState fm_state(other_data);
+	RowOperationsState row_state(*aggregate_allocator);
+	while (fm_state.Scan()) {
+		FindOrCreateGroups(fm_state.groups, fm_state.hashes, fm_state.group_addresses, fm_state.new_groups_sel);
+		RowOperations::CombineStates(row_state, layout, fm_state.scan_state.chunk_state.row_locations,
+		                             fm_state.group_addresses, fm_state.groups.size());
+		if (layout.HasDestructor()) {
+			RowOperations::DestroyStates(row_state, layout, fm_state.scan_state.chunk_state.row_locations,
+			                             fm_state.groups.size());
+		}
 	}
 
 	Verify();
 }
 
-void GroupedAggregateHashTable::Partition(vector<GroupedAggregateHashTable *> &partition_hts, idx_t radix_bits) {
-	const auto num_partitions = RadixPartitioning::NumberOfPartitions(radix_bits);
-	D_ASSERT(partition_hts.size() == num_partitions);
-
-	// Partition the data
-	auto partitioned_data =
-	    make_uniq<RadixPartitionedTupleData>(buffer_manager, layout, radix_bits, layout.ColumnCount() - 1);
-	partitioned_data->Partition(*data_collection, TupleDataPinProperties::KEEP_EVERYTHING_PINNED);
-	D_ASSERT(partitioned_data->GetPartitions().size() == num_partitions);
-
-	// Move the partitioned data collections to the partitioned hash tables and initialize the 1st part of the HT
-	auto &partitions = partitioned_data->GetPartitions();
-	for (idx_t partition_idx = 0; partition_idx < num_partitions; partition_idx++) {
-		auto &partition_ht = *partition_hts[partition_idx];
-		partition_ht.data_collection = std::move(partitions[partition_idx]);
-		partition_ht.aggregate_allocator = aggregate_allocator;
-		partition_ht.InitializeFirstPart();
-		partition_ht.Verify();
-	}
-}
-
-void GroupedAggregateHashTable::InitializeFirstPart() {
-	data_collection->GetBlockPointers(payload_hds_ptrs);
-	auto size = MaxValue<idx_t>(NextPowerOfTwo(Count() * 2L), capacity);
-	switch (entry_type) {
-	case HtEntryType::HT_WIDTH_64:
-		Resize<aggr_ht_entry_64>(size);
-		break;
-	case HtEntryType::HT_WIDTH_32:
-		Resize<aggr_ht_entry_32>(size);
-		break;
-	default:
-		throw InternalException("Unknown HT entry width");
-	}
-}
-
-idx_t GroupedAggregateHashTable::Scan(TupleDataParallelScanState &gstate, TupleDataLocalScanState &lstate,
-                                      DataChunk &result) {
-	data_collection->Scan(gstate, lstate, result);
-
-	RowOperationsState row_state(aggregate_allocator->GetAllocator());
-	const auto group_cols = layout.ColumnCount() - 1;
-	RowOperations::FinalizeStates(row_state, layout, lstate.chunk_state.row_locations, result, group_cols);
-
-	return result.size();
-}
-
-void GroupedAggregateHashTable::Finalize() {
-	if (is_finalized) {
-		return;
-	}
-
-	// Early release hashes (not needed for partition/scan) and data collection (will be pinned again when scanning)
-	hashes_hdl.Destroy();
-	data_collection->FinalizePinState(td_pin_state);
-	data_collection->Unpin();
-
-	is_finalized = true;
+void GroupedAggregateHashTable::UnpinData() {
+	partitioned_data->FlushAppendState(state.append_state);
+	partitioned_data->Unpin();
 }
 
 } // namespace duckdb