duckdb 0.8.2-dev157.0 → 0.8.2-dev1573.0

package/src/duckdb/src/storage/serialization/serialize_tableref.cpp

@@ -0,0 +1,166 @@
+//===----------------------------------------------------------------------===//
+// This file is automatically generated by scripts/generate_serialization.py
+// Do not edit this file manually, your changes will be overwritten
+//===----------------------------------------------------------------------===//
+
+#include "duckdb/common/serializer/format_serializer.hpp"
+#include "duckdb/common/serializer/format_deserializer.hpp"
+#include "duckdb/parser/tableref/list.hpp"
+
+namespace duckdb {
+
+void TableRef::FormatSerialize(FormatSerializer &serializer) const {
+	serializer.WriteProperty("type", type);
+	serializer.WriteProperty("alias", alias);
+	serializer.WriteOptionalProperty("sample", sample);
+}
+
+unique_ptr<TableRef> TableRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto type = deserializer.ReadProperty<TableReferenceType>("type");
+	auto alias = deserializer.ReadProperty<string>("alias");
+	auto sample = deserializer.ReadOptionalProperty<unique_ptr<SampleOptions>>("sample");
+	unique_ptr<TableRef> result;
+	switch (type) {
+	case TableReferenceType::BASE_TABLE:
+		result = BaseTableRef::FormatDeserialize(deserializer);
+		break;
+	case TableReferenceType::EMPTY:
+		result = EmptyTableRef::FormatDeserialize(deserializer);
+		break;
+	case TableReferenceType::EXPRESSION_LIST:
+		result = ExpressionListRef::FormatDeserialize(deserializer);
+		break;
+	case TableReferenceType::JOIN:
+		result = JoinRef::FormatDeserialize(deserializer);
+		break;
+	case TableReferenceType::PIVOT:
+		result = PivotRef::FormatDeserialize(deserializer);
+		break;
+	case TableReferenceType::SUBQUERY:
+		result = SubqueryRef::FormatDeserialize(deserializer);
+		break;
+	case TableReferenceType::TABLE_FUNCTION:
+		result = TableFunctionRef::FormatDeserialize(deserializer);
+		break;
+	default:
+		throw SerializationException("Unsupported type for deserialization of TableRef!");
+	}
+	result->alias = std::move(alias);
+	result->sample = std::move(sample);
+	return result;
+}
+
+void BaseTableRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+	serializer.WriteProperty("schema_name", schema_name);
+	serializer.WriteProperty("table_name", table_name);
+	serializer.WriteProperty("column_name_alias", column_name_alias);
+	serializer.WriteProperty("catalog_name", catalog_name);
+}
+
+unique_ptr<TableRef> BaseTableRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<BaseTableRef>(new BaseTableRef());
+	deserializer.ReadProperty("schema_name", result->schema_name);
+	deserializer.ReadProperty("table_name", result->table_name);
+	deserializer.ReadProperty("column_name_alias", result->column_name_alias);
+	deserializer.ReadProperty("catalog_name", result->catalog_name);
+	return std::move(result);
+}
+
+void JoinRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+	serializer.WriteProperty("left", *left);
+	serializer.WriteProperty("right", *right);
+	serializer.WriteOptionalProperty("condition", condition);
+	serializer.WriteProperty("type", type);
+	serializer.WriteProperty("ref_type", ref_type);
+	serializer.WriteProperty("using_columns", using_columns);
+}
+
+unique_ptr<TableRef> JoinRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<JoinRef>(new JoinRef());
+	deserializer.ReadProperty("left", result->left);
+	deserializer.ReadProperty("right", result->right);
+	deserializer.ReadOptionalProperty("condition", result->condition);
+	deserializer.ReadProperty("type", result->type);
+	deserializer.ReadProperty("ref_type", result->ref_type);
+	deserializer.ReadProperty("using_columns", result->using_columns);
+	return std::move(result);
+}
+
+void SubqueryRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+	serializer.WriteProperty("subquery", *subquery);
+	serializer.WriteProperty("column_name_alias", column_name_alias);
+}
+
+unique_ptr<TableRef> SubqueryRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<SubqueryRef>(new SubqueryRef());
+	deserializer.ReadProperty("subquery", result->subquery);
+	deserializer.ReadProperty("column_name_alias", result->column_name_alias);
+	return std::move(result);
+}
+
+void TableFunctionRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+	serializer.WriteProperty("function", *function);
+	serializer.WriteProperty("alias", alias);
+	serializer.WriteProperty("column_name_alias", column_name_alias);
+}
+
+unique_ptr<TableRef> TableFunctionRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<TableFunctionRef>(new TableFunctionRef());
+	deserializer.ReadProperty("function", result->function);
+	deserializer.ReadProperty("alias", result->alias);
+	deserializer.ReadProperty("column_name_alias", result->column_name_alias);
+	return std::move(result);
+}
+
+void EmptyTableRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+}
+
+unique_ptr<TableRef> EmptyTableRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<EmptyTableRef>(new EmptyTableRef());
+	return std::move(result);
+}
+
+void ExpressionListRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+	serializer.WriteProperty("expected_names", expected_names);
+	serializer.WriteProperty("expected_types", expected_types);
+	serializer.WriteProperty("values", values);
+}
+
+unique_ptr<TableRef> ExpressionListRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<ExpressionListRef>(new ExpressionListRef());
+	deserializer.ReadProperty("expected_names", result->expected_names);
+	deserializer.ReadProperty("expected_types", result->expected_types);
+	deserializer.ReadProperty("values", result->values);
+	return std::move(result);
+}
+
+void PivotRef::FormatSerialize(FormatSerializer &serializer) const {
+	TableRef::FormatSerialize(serializer);
+	serializer.WriteProperty("source", *source);
+	serializer.WriteProperty("aggregates", aggregates);
+	serializer.WriteProperty("unpivot_names", unpivot_names);
+	serializer.WriteProperty("pivots", pivots);
+	serializer.WriteProperty("groups", groups);
+	serializer.WriteProperty("column_name_alias", column_name_alias);
+	serializer.WriteProperty("include_nulls", include_nulls);
+}
+
+unique_ptr<TableRef> PivotRef::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::unique_ptr<PivotRef>(new PivotRef());
+	deserializer.ReadProperty("source", result->source);
+	deserializer.ReadProperty("aggregates", result->aggregates);
+	deserializer.ReadProperty("unpivot_names", result->unpivot_names);
+	deserializer.ReadProperty("pivots", result->pivots);
+	deserializer.ReadProperty("groups", result->groups);
+	deserializer.ReadProperty("column_name_alias", result->column_name_alias);
+	deserializer.ReadProperty("include_nulls", result->include_nulls);
+	return std::move(result);
+}
+
+} // namespace duckdb

package/src/duckdb/src/storage/serialization/serialize_types.cpp

@@ -0,0 +1,127 @@
+//===----------------------------------------------------------------------===//
+// This file is automatically generated by scripts/generate_serialization.py
+// Do not edit this file manually, your changes will be overwritten
+//===----------------------------------------------------------------------===//
+
+#include "duckdb/common/serializer/format_serializer.hpp"
+#include "duckdb/common/serializer/format_deserializer.hpp"
+#include "duckdb/common/extra_type_info.hpp"
+
+namespace duckdb {
+
+void ExtraTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	serializer.WriteProperty("type", type);
+	serializer.WriteProperty("alias", alias);
+}
+
+shared_ptr<ExtraTypeInfo> ExtraTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto type = deserializer.ReadProperty<ExtraTypeInfoType>("type");
+	auto alias = deserializer.ReadProperty<string>("alias");
+	shared_ptr<ExtraTypeInfo> result;
+	switch (type) {
+	case ExtraTypeInfoType::AGGREGATE_STATE_TYPE_INFO:
+		result = AggregateStateTypeInfo::FormatDeserialize(deserializer);
+		break;
+	case ExtraTypeInfoType::DECIMAL_TYPE_INFO:
+		result = DecimalTypeInfo::FormatDeserialize(deserializer);
+		break;
+	case ExtraTypeInfoType::ENUM_TYPE_INFO:
+		result = EnumTypeInfo::FormatDeserialize(deserializer);
+		break;
+	case ExtraTypeInfoType::GENERIC_TYPE_INFO:
+		result = make_shared<ExtraTypeInfo>(type);
+		break;
+	case ExtraTypeInfoType::INVALID_TYPE_INFO:
+		return nullptr;
+	case ExtraTypeInfoType::LIST_TYPE_INFO:
+		result = ListTypeInfo::FormatDeserialize(deserializer);
+		break;
+	case ExtraTypeInfoType::STRING_TYPE_INFO:
+		result = StringTypeInfo::FormatDeserialize(deserializer);
+		break;
+	case ExtraTypeInfoType::STRUCT_TYPE_INFO:
+		result = StructTypeInfo::FormatDeserialize(deserializer);
+		break;
+	case ExtraTypeInfoType::USER_TYPE_INFO:
+		result = UserTypeInfo::FormatDeserialize(deserializer);
+		break;
+	default:
+		throw SerializationException("Unsupported type for deserialization of ExtraTypeInfo!");
+	}
+	result->alias = std::move(alias);
+	return result;
+}
+
+void DecimalTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	ExtraTypeInfo::FormatSerialize(serializer);
+	serializer.WriteProperty("width", width);
+	serializer.WriteProperty("scale", scale);
+}
+
+shared_ptr<ExtraTypeInfo> DecimalTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::shared_ptr<DecimalTypeInfo>(new DecimalTypeInfo());
+	deserializer.ReadProperty("width", result->width);
+	deserializer.ReadProperty("scale", result->scale);
+	return std::move(result);
+}
+
+void StringTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	ExtraTypeInfo::FormatSerialize(serializer);
+	serializer.WriteProperty("collation", collation);
+}
+
+shared_ptr<ExtraTypeInfo> StringTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::shared_ptr<StringTypeInfo>(new StringTypeInfo());
+	deserializer.ReadProperty("collation", result->collation);
+	return std::move(result);
+}
+
+void ListTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	ExtraTypeInfo::FormatSerialize(serializer);
+	serializer.WriteProperty("child_type", child_type);
+}
+
+shared_ptr<ExtraTypeInfo> ListTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::shared_ptr<ListTypeInfo>(new ListTypeInfo());
+	deserializer.ReadProperty("child_type", result->child_type);
+	return std::move(result);
+}
+
+void StructTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	ExtraTypeInfo::FormatSerialize(serializer);
+	serializer.WriteProperty("child_types", child_types);
+}
+
+shared_ptr<ExtraTypeInfo> StructTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::shared_ptr<StructTypeInfo>(new StructTypeInfo());
+	deserializer.ReadProperty("child_types", result->child_types);
+	return std::move(result);
+}
+
+void AggregateStateTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	ExtraTypeInfo::FormatSerialize(serializer);
+	serializer.WriteProperty("function_name", state_type.function_name);
+	serializer.WriteProperty("return_type", state_type.return_type);
+	serializer.WriteProperty("bound_argument_types", state_type.bound_argument_types);
+}
+
+shared_ptr<ExtraTypeInfo> AggregateStateTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::shared_ptr<AggregateStateTypeInfo>(new AggregateStateTypeInfo());
+	deserializer.ReadProperty("function_name", result->state_type.function_name);
+	deserializer.ReadProperty("return_type", result->state_type.return_type);
+	deserializer.ReadProperty("bound_argument_types", result->state_type.bound_argument_types);
+	return std::move(result);
+}
+
+void UserTypeInfo::FormatSerialize(FormatSerializer &serializer) const {
+	ExtraTypeInfo::FormatSerialize(serializer);
+	serializer.WriteProperty("user_type_name", user_type_name);
+}
+
+shared_ptr<ExtraTypeInfo> UserTypeInfo::FormatDeserialize(FormatDeserializer &deserializer) {
+	auto result = duckdb::shared_ptr<UserTypeInfo>(new UserTypeInfo());
+	deserializer.ReadProperty("user_type_name", result->user_type_name);
+	return std::move(result);
+}
+
+} // namespace duckdb

package/src/duckdb/src/storage/single_file_block_manager.cpp

@@ -368,6 +368,29 @@ void SingleFileBlockManager::Write(FileBuffer &buffer, block_id_t block_id) {
 	ChecksumAndWrite(buffer, BLOCK_START + block_id * Storage::BLOCK_ALLOC_SIZE);
 }
 
+void SingleFileBlockManager::Truncate() {
+	BlockManager::Truncate();
+	idx_t blocks_to_truncate = 0;
+	// reverse iterate over the free-list
+	for (auto entry = free_list.rbegin(); entry != free_list.rend(); entry++) {
+		auto block_id = *entry;
+		if (block_id + 1 != max_block) {
+			break;
+		}
+		blocks_to_truncate++;
+		max_block--;
+	}
+	if (blocks_to_truncate == 0) {
+		// nothing to truncate
+		return;
+	}
+	// truncate the file
+	for (idx_t i = 0; i < blocks_to_truncate; i++) {
+		free_list.erase(max_block + i);
+	}
+	handle->Truncate(BLOCK_START + max_block * Storage::BLOCK_ALLOC_SIZE);
+}
+
 vector<block_id_t> SingleFileBlockManager::GetFreeListBlocks() {
 	vector<block_id_t> free_list_blocks;
 

package/src/duckdb/src/storage/statistics/string_stats.cpp

@@ -1,10 +1,11 @@
 #include "duckdb/storage/statistics/string_stats.hpp"
-#include "duckdb/storage/statistics/base_statistics.hpp"
+
 #include "duckdb/common/field_writer.hpp"
-#include "utf8proc_wrapper.hpp"
 #include "duckdb/common/string_util.hpp"
 #include "duckdb/common/types/vector.hpp"
 #include "duckdb/main/error_manager.hpp"
+#include "duckdb/storage/statistics/base_statistics.hpp"
+#include "utf8proc_wrapper.hpp"
 
 namespace duckdb {
 
@@ -67,6 +68,24 @@ bool StringStats::CanContainUnicode(const BaseStatistics &stats) {
 	return StringStats::GetDataUnsafe(stats).has_unicode;
 }
 
+string GetStringMinMaxValue(const data_t data[]) {
+	idx_t len;
+	for (len = 0; len < StringStatsData::MAX_STRING_MINMAX_SIZE; len++) {
+		if (!data[len]) {
+			break;
+		}
+	}
+	return string(const_char_ptr_cast(data), len);
+}
+
+string StringStats::Min(const BaseStatistics &stats) {
+	return GetStringMinMaxValue(StringStats::GetDataUnsafe(stats).min);
+}
+
+string StringStats::Max(const BaseStatistics &stats) {
+	return GetStringMinMaxValue(StringStats::GetDataUnsafe(stats).max);
+}
+
 void StringStats::ResetMaxStringLength(BaseStatistics &stats) {
 	StringStats::GetDataUnsafe(stats).has_max_string_length = false;
 }

package/src/duckdb/src/storage/storage_info.cpp

@@ -2,7 +2,7 @@
 
 namespace duckdb {
 
-const uint64_t VERSION_NUMBER = 51;
+const uint64_t VERSION_NUMBER = 52;
 
 struct StorageVersionInfo {
 	const char *version_name;

package/src/duckdb/src/storage/table/chunk_info.cpp

@@ -89,6 +89,10 @@ void ChunkConstantInfo::Serialize(Serializer &serializer) {
 	serializer.Write<idx_t>(start);
 }
 
+idx_t ChunkConstantInfo::GetCommittedDeletedCount(idx_t max_count) {
+	return delete_id < TRANSACTION_ID_START ? max_count : 0;
+}
+
 unique_ptr<ChunkInfo> ChunkConstantInfo::Deserialize(Deserializer &source) {
 	auto start = source.Read<idx_t>();
 
@@ -244,6 +248,19 @@ void ChunkVectorInfo::Serialize(Serializer &serializer) {
 	serializer.WriteData(data_ptr_cast(deleted_tuples), sizeof(bool) * STANDARD_VECTOR_SIZE);
 }
 
+idx_t ChunkVectorInfo::GetCommittedDeletedCount(idx_t max_count) {
+	if (!any_deleted) {
+		return 0;
+	}
+	idx_t delete_count = 0;
+	for (idx_t i = 0; i < max_count; i++) {
+		if (deleted[i] < TRANSACTION_ID_START) {
+			delete_count++;
+		}
+	}
+	return delete_count;
+}
+
 unique_ptr<ChunkInfo> ChunkVectorInfo::Deserialize(Deserializer &source) {
 	auto start = source.Read<idx_t>();
 

package/src/duckdb/src/storage/table/row_group.cpp

@@ -67,6 +67,21 @@ void VersionNode::SetStart(idx_t start) {
 	}
 }
 
+idx_t VersionNode::GetCommittedDeletedCount(idx_t count) {
+	idx_t deleted_count = 0;
+	for (idx_t r = 0, i = 0; r < count; r += STANDARD_VECTOR_SIZE, i++) {
+		if (!info[i]) {
+			continue;
+		}
+		idx_t max_count = MinValue<idx_t>(STANDARD_VECTOR_SIZE, count - r);
+		if (max_count == 0) {
+			break;
+		}
+		deleted_count += info[i]->GetCommittedDeletedCount(max_count);
+	}
+	return deleted_count;
+}
+
 RowGroup::~RowGroup() {
 }
 
@@ -112,10 +127,6 @@ ColumnData &RowGroup::GetColumn(storage_t c) {
 	return *columns[c];
 }
 
-DatabaseInstance &RowGroup::GetDatabase() {
-	return GetCollection().GetDatabase();
-}
-
 BlockManager &RowGroup::GetBlockManager() {
 	return GetCollection().GetBlockManager();
 }
@@ -270,7 +281,7 @@ unique_ptr<RowGroup> RowGroup::AlterType(RowGroupCollection &new_collection, con
 }
 
 unique_ptr<RowGroup> RowGroup::AddColumn(RowGroupCollection &new_collection, ColumnDefinition &new_column,
-                                         ExpressionExecutor &executor, Expression *default_value, Vector &result) {
+                                         ExpressionExecutor &executor, Expression &default_value, Vector &result) {
 	Verify();
 
 	// construct a new column data for the new column
@@ -285,10 +296,8 @@ unique_ptr<RowGroup> RowGroup::AddColumn(RowGroupCollection &new_collection, Col
 		added_column->InitializeAppend(state);
 		for (idx_t i = 0; i < rows_to_write; i += STANDARD_VECTOR_SIZE) {
 			idx_t rows_in_this_vector = MinValue<idx_t>(rows_to_write - i, STANDARD_VECTOR_SIZE);
-			if (default_value) {
-				dummy_chunk.SetCardinality(rows_in_this_vector);
-				executor.ExecuteExpression(dummy_chunk, result);
-			}
+			dummy_chunk.SetCardinality(rows_in_this_vector);
+			executor.ExecuteExpression(dummy_chunk, result);
 			added_column->Append(state, result, rows_in_this_vector);
 		}
 	}
@@ -796,6 +805,13 @@ RowGroupWriteData RowGroup::WriteToDisk(PartialBlockManager &manager,
 	return result;
 }
 
+bool RowGroup::AllDeleted() {
+	if (!version_info) {
+		return false;
+	}
+	return version_info->GetCommittedDeletedCount(count) == count;
+}
+
 RowGroupPointer RowGroup::Checkpoint(RowGroupWriter &writer, TableStatistics &global_stats) {
 	RowGroupPointer row_group_pointer;
 

package/src/duckdb/src/storage/table/row_group_collection.cpp

@@ -67,10 +67,6 @@ AttachedDatabase &RowGroupCollection::GetAttached() {
 	return GetTableInfo().db;
 }
 
-DatabaseInstance &RowGroupCollection::GetDatabase() {
-	return GetAttached().GetDatabase();
-}
-
 //===--------------------------------------------------------------------===//
 // Initialize
 //===--------------------------------------------------------------------===//
@@ -97,10 +93,6 @@ RowGroup *RowGroupCollection::GetRowGroup(int64_t index) {
 	return (RowGroup *)row_groups->GetSegmentByIndex(index);
 }
 
-idx_t RowGroupCollection::RowGroupCount() {
-	return row_groups->GetSegmentCount();
-}
-
 void RowGroupCollection::Verify() {
 #ifdef DEBUG
 	idx_t current_total_rows = 0;
@@ -600,11 +592,24 @@ void RowGroupCollection::UpdateColumn(TransactionData transaction, Vector &row_i
 // Checkpoint
 //===--------------------------------------------------------------------===//
 void RowGroupCollection::Checkpoint(TableDataWriter &writer, TableStatistics &global_stats) {
-	for (auto &row_group : row_groups->Segments()) {
-		auto rowg_writer = writer.GetRowGroupWriter(row_group);
-		auto pointer = row_group.Checkpoint(*rowg_writer, global_stats);
-		writer.AddRowGroup(std::move(pointer), std::move(rowg_writer));
+	bool can_vacuum_deletes = info->indexes.Empty();
+	idx_t start = this->row_start;
+	auto segments = row_groups->MoveSegments();
+	auto l = row_groups->Lock();
+	for (auto &entry : segments) {
+		auto &row_group = *entry.node;
+		if (can_vacuum_deletes && row_group.AllDeleted()) {
+			row_group.CommitDrop();
+			continue;
+		}
+		row_group.MoveToCollection(*this, start);
+		auto row_group_writer = writer.GetRowGroupWriter(row_group);
+		auto pointer = row_group.Checkpoint(*row_group_writer, global_stats);
+		writer.AddRowGroup(std::move(pointer), std::move(row_group_writer));
+		row_groups->AppendSegment(l, std::move(entry.node));
+		start += row_group.count;
 	}
+	total_rows = start;
 }
 
 //===--------------------------------------------------------------------===//
@@ -637,7 +642,7 @@ vector<ColumnSegmentInfo> RowGroupCollection::GetColumnSegmentInfo() {
 // Alter
 //===--------------------------------------------------------------------===//
 shared_ptr<RowGroupCollection> RowGroupCollection::AddColumn(ClientContext &context, ColumnDefinition &new_column,
-                                                             Expression *default_value) {
+                                                             Expression &default_value) {
 	idx_t new_column_idx = types.size();
 	auto new_types = types;
 	new_types.push_back(new_column.GetType());
@@ -647,11 +652,7 @@ shared_ptr<RowGroupCollection> RowGroupCollection::AddColumn(ClientContext &cont
 	ExpressionExecutor executor(context);
 	DataChunk dummy_chunk;
 	Vector default_vector(new_column.GetType());
-	if (!default_value) {
-		FlatVector::Validity(default_vector).SetAllInvalid(STANDARD_VECTOR_SIZE);
-	} else {
-		executor.AddExpression(*default_value);
-	}
+	executor.AddExpression(default_value);
 
 	result->stats.InitializeAddColumn(stats, new_column.GetType());
 	auto &new_column_stats = result->stats.GetStats(new_column_idx);

package/src/duckdb/third_party/concurrentqueue/concurrentqueue.h

@@ -80,7 +80,7 @@ namespace duckdb_moodycamel { namespace details {
 	static const thread_id_t invalid_thread_id2 = 0xFFFFFFFFU;	// Not technically guaranteed to be invalid, but is never used in practice. Note that all Win32 thread IDs are presently multiples of 4.
 	static inline thread_id_t thread_id() { return static_cast<thread_id_t>(::GetCurrentThreadId()); }
 } }
-#elif defined(__arm__) || defined(_M_ARM) || defined(__aarch64__) || (defined(__APPLE__) && TARGET_OS_IPHONE)
+#elif defined(__arm__) || defined(_M_ARM) || defined(__aarch64__) || (defined(__APPLE__) && TARGET_OS_IPHONE) || defined(__MVS__)
 namespace duckdb_moodycamel { namespace details {
 	static_assert(sizeof(std::thread::id) == 4 || sizeof(std::thread::id) == 8, "std::thread::id is expected to be either 4 or 8 bytes");
 	
@@ -194,7 +194,7 @@ namespace duckdb_moodycamel { namespace details {
 // VS2013 doesn't support `thread_local`, and MinGW-w64 w/ POSIX threading has a crippling bug: http://sourceforge.net/p/mingw-w64/bugs/445
 // g++ <=4.7 doesn't support thread_local either.
 // Finally, iOS/ARM doesn't have support for it either, and g++/ARM allows it to compile but it's unconfirmed to actually work
-#if (!defined(_MSC_VER) || _MSC_VER >= 1900) && (!defined(__MINGW32__) && !defined(__MINGW64__) || !defined(__WINPTHREADS_VERSION)) && (!defined(__GNUC__) || __GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)) && (!defined(__APPLE__) || !TARGET_OS_IPHONE) && !defined(__arm__) && !defined(_M_ARM) && !defined(__aarch64__)
+#if (!defined(_MSC_VER) || _MSC_VER >= 1900) && (!defined(__MINGW32__) && !defined(__MINGW64__) || !defined(__WINPTHREADS_VERSION)) && (!defined(__GNUC__) || __GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 8)) && (!defined(__APPLE__) || !TARGET_OS_IPHONE) && !defined(__arm__) && !defined(_M_ARM) && !defined(__aarch64__) && !defined(__MVS__)
 // Assume `thread_local` is fully supported in all other C++11 compilers/platforms
 //#define MOODYCAMEL_CPP11_THREAD_LOCAL_SUPPORTED    // always disabled for now since several users report having problems with it on
 #endif

package/src/duckdb/third_party/concurrentqueue/lightweightsemaphore.h

@@ -27,6 +27,8 @@ extern "C" {
 #elif defined(__unix__)
 #include <semaphore.h>
 #include <chrono>
+#elif defined(__MVS__)
+#include <zos-semaphore.h>
 #endif
 
 namespace duckdb_moodycamel
@@ -254,6 +256,80 @@ public:
 		}
 	}
 };
+#elif defined(__MVS__)
+//---------------------------------------------------------
+// Semaphore (MVS aka z/OS)
+//---------------------------------------------------------
+class Semaphore
+{
+private:
+	sem_t m_sema;
+
+	Semaphore(const Semaphore& other) MOODYCAMEL_DELETE_FUNCTION;
+	Semaphore& operator=(const Semaphore& other) MOODYCAMEL_DELETE_FUNCTION;
+
+public:
+	Semaphore(int initialCount = 0)
+	{
+		assert(initialCount >= 0);
+		int rc = sem_init(&m_sema, 0, initialCount);
+		assert(rc == 0);
+		(void)rc;
+	}
+
+	~Semaphore()
+	{
+		sem_destroy(&m_sema);
+	}
+
+	bool wait()
+	{
+		// http://stackoverflow.com/questions/2013181/gdb-causes-sem-wait-to-fail-with-eintr-error
+		int rc;
+		do {
+			rc = sem_wait(&m_sema);
+		} while (rc == -1 && errno == EINTR);
+		return rc == 0;
+	}
+
+	bool try_wait()
+	{
+		int rc;
+		do {
+			rc = sem_trywait(&m_sema);
+		} while (rc == -1 && errno == EINTR);
+		return rc == 0;
+	}
+
+	bool timed_wait(std::uint64_t usecs)
+	{
+		struct timespec ts;
+		const int usecs_in_1_sec = 1000000;
+		const int nsecs_in_1_sec = 1000000000;
+
+		ts.tv_sec = usecs / usecs_in_1_sec;
+		ts.tv_nsec = (usecs % usecs_in_1_sec) * 1000;
+
+		int rc;
+		do {
+			rc = sem_timedwait(&m_sema, &ts);
+		} while (rc == -1 && errno == EINTR);
+		return rc == 0;
+	}
+
+	void signal()
+	{
+		while (sem_post(&m_sema) == -1);
+	}
+
+	void signal(int count)
+	{
+		while (count-- > 0)
+		{
+			while (sem_post(&m_sema) == -1);
+		}
+	}
+};
 #else
 #error Unsupported platform! (No semaphore wrapper available)
 #endif

package/src/duckdb/third_party/fast_float/fast_float/fast_float.h

@@ -130,6 +130,8 @@ from_chars_result from_chars(const char *first, const char *last,
 #include <machine/endian.h>
 #elif defined(sun) || defined(__sun)
 #include <sys/byteorder.h>
+#elif defined(__MVS__)
+#include <sys/endian.h>
 #else
 #include <endian.h>
 #endif