duckdb 0.7.2-dev2867.0 → 0.7.2-dev3117.0

package/binding.gyp

@@ -135,7 +135,9 @@
                 "src/duckdb/ub_src_transaction.cpp", 
                 "src/duckdb/src/verification/copied_statement_verifier.cpp", 
                 "src/duckdb/src/verification/deserialized_statement_verifier.cpp", 
+                "src/duckdb/src/verification/deserialized_statement_verifier_v2.cpp", 
                 "src/duckdb/src/verification/external_statement_verifier.cpp", 
+                "src/duckdb/src/verification/no_operator_caching_verifier.cpp", 
                 "src/duckdb/src/verification/parsed_statement_verifier.cpp", 
                 "src/duckdb/src/verification/prepared_statement_verifier.cpp", 
                 "src/duckdb/src/verification/statement_verifier.cpp", 

package/package.json

@@ -2,7 +2,7 @@
   "name": "duckdb",
   "main": "./lib/duckdb.js",
   "types": "./lib/duckdb.d.ts",
-  "version": "0.7.2-dev2867.0",
+  "version": "0.7.2-dev3117.0",
   "description": "DuckDB node.js API",
   "gypfile": true,
   "dependencies": {

package/src/duckdb/extension/icu/icu-datepart.cpp

@@ -204,7 +204,11 @@ struct ICUDatePart : public ICUDateFunc {
 	}
 
 	static string_t MonthName(icu::Calendar *calendar, const uint64_t micros) {
-		return Date::MONTH_NAMES[ExtractMonth(calendar, micros) - 1];
+		const auto mm = ExtractMonth(calendar, micros) - 1;
+		if (mm == 12) {
+			return "Undecember";
+		}
+		return Date::MONTH_NAMES[mm];
 	}
 
 	static string_t DayName(icu::Calendar *calendar, const uint64_t micros) {

package/src/duckdb/extension/json/include/json_deserializer.hpp

@@ -7,6 +7,7 @@ namespace duckdb {
 class JsonDeserializer : public FormatDeserializer {
 public:
 	JsonDeserializer(yyjson_val *val, yyjson_doc *doc) : doc(doc) {
+		deserialize_enum_from_string = true;
 		stack.emplace_back(val);
 	}
 	~JsonDeserializer() {

package/src/duckdb/extension/json/include/json_serializer.hpp

@@ -24,12 +24,19 @@ private:
 	// Either adds a value to the current object with the current tag, or appends it to the current array
 	void PushValue(yyjson_mut_val *val);
 
-public:
 	explicit JsonSerializer(yyjson_mut_doc *doc, bool skip_if_null, bool skip_if_empty)
 	    : doc(doc), stack({yyjson_mut_obj(doc)}), skip_if_null(skip_if_null), skip_if_empty(skip_if_empty) {
 		serialize_enum_as_string = true;
 	}
 
+public:
+	template <class T>
+	static yyjson_mut_val *Serialize(T &value, yyjson_mut_doc *doc, bool skip_if_null, bool skip_if_empty) {
+		JsonSerializer serializer(doc, skip_if_null, skip_if_empty);
+		value.FormatSerialize(serializer);
+		return serializer.GetRootObject();
+	}
+
 	yyjson_mut_val *GetRootObject() {
 		D_ASSERT(stack.size() == 1); // or we forgot to pop somewhere
 		return stack.front();

package/src/duckdb/extension/json/json_functions/json_serialize_sql.cpp

@@ -98,9 +98,7 @@ static void JsonSerializeFunction(DataChunk &args, ExpressionState &state, Vecto
 					throw NotImplementedException("Only SELECT statements can be serialized to json!");
 				}
 				auto &select = statement->Cast<SelectStatement>();
-				auto serializer = JsonSerializer(doc, info.skip_if_null, info.skip_if_empty);
-				select.FormatSerialize(serializer);
-				auto json = serializer.GetRootObject();
+				auto json = JsonSerializer::Serialize(select, doc, info.skip_if_null, info.skip_if_empty);
 
 				yyjson_mut_arr_append(statements_arr, json);
 			}

package/src/duckdb/extension/json/json_functions/json_structure.cpp

@@ -3,6 +3,7 @@
 #include "json_executors.hpp"
 #include "json_scan.hpp"
 #include "json_transform.hpp"
+#include "duckdb/common/enum_util.hpp"
 
 namespace duckdb {
 
@@ -300,7 +301,7 @@ bool JSONStructureNode::EliminateCandidateFormats(idx_t count, Vector &string_ve
 			success = TryParse<TryParseTimeStamp, timestamp_t>(string_vector, format, count);
 			break;
 		default:
-			throw InternalException("No date/timestamp formats for %s", LogicalTypeIdToString(type));
+			throw InternalException("No date/timestamp formats for %s", EnumUtil::ToString(type));
 		}
 		if (success) {
 			while (formats.size() > i) {
@@ -457,8 +458,7 @@ static inline yyjson_mut_val *ConvertStructure(const JSONStructureNode &node, yy
 	case LogicalTypeId::STRUCT:
 		return ConvertStructureObject(node, doc);
 	default:
-		auto type_string = LogicalTypeIdToString(desc.type); // TODO: this requires copying, can be optimized
-		return yyjson_mut_strncpy(doc, type_string.c_str(), type_string.length());
+		return yyjson_mut_str(doc, EnumUtil::ToChars(desc.type));
 	}
 }
 

package/src/duckdb/extension/json/json_functions/json_transform.cpp

@@ -1,6 +1,7 @@
 #include "json_transform.hpp"
 
 #include "duckdb/common/types.hpp"
+#include "duckdb/common/enum_util.hpp"
 #include "duckdb/execution/expression_executor.hpp"
 #include "duckdb/function/scalar/nested_functions.hpp"
 #include "json_functions.hpp"
@@ -266,7 +267,7 @@ bool JSONTransform::GetStringVector(yyjson_val *vals[], const idx_t count, const
 		if (!val || unsafe_yyjson_is_null(val)) {
 			validity.SetInvalid(i);
 		} else if (options.strict_cast && !unsafe_yyjson_is_str(val)) {
-			options.error_message = StringUtil::Format("Unable to cast '%s' to " + LogicalTypeIdToString(target.id()),
+			options.error_message = StringUtil::Format("Unable to cast '%s' to " + EnumUtil::ToString(target.id()),
 			                                           JSONCommon::ValToString(val, 50));
 			options.object_index = i;
 			return false;
@@ -343,7 +344,7 @@ static bool TransformFromStringWithFormat(yyjson_val *vals[], Vector &result, co
 	case LogicalTypeId::TIMESTAMP:
 		return TransformStringWithFormat<TryParseTimeStamp, timestamp_t>(string_vector, format, count, result, options);
 	default:
-		throw InternalException("No date/timestamp formats for %s", LogicalTypeIdToString(result.GetType().id()));
+		throw InternalException("No date/timestamp formats for %s", EnumUtil::ToString(result.GetType().id()));
 	}
 }
 

package/src/duckdb/extension/parquet/parquet-extension.cpp

@@ -304,6 +304,9 @@ public:
 				parquet_options.file_row_number = BooleanValue::Get(kv.second);
 			}
 		}
+		if (parquet_options.file_options.auto_detect_hive_partitioning) {
+			parquet_options.file_options.hive_partitioning = MultiFileReaderOptions::AutoDetectHivePartitioning(files);
+		}
 		return ParquetScanBindInternal(context, std::move(files), return_types, names, parquet_options);
 	}
 
@@ -665,14 +668,13 @@ unique_ptr<LocalFunctionData> ParquetWriteInitializeLocal(ExecutionContext &cont
 }
 
 //===--------------------------------------------------------------------===//
-// Parallel
+// Execution Mode
 //===--------------------------------------------------------------------===//
-bool ParquetWriteIsParallel(ClientContext &context, FunctionData &bind_data) {
-	auto &config = DBConfig::GetConfig(context);
-	if (config.options.preserve_insertion_order) {
-		return false;
+CopyFunctionExecutionMode ParquetWriteExecutionMode(bool preserve_insertion_order, bool supports_batch_index) {
+	if (!preserve_insertion_order) {
+		return CopyFunctionExecutionMode::PARALLEL_COPY_TO_FILE;
 	}
-	return true;
+	return CopyFunctionExecutionMode::REGULAR_COPY_TO_FILE;
 }
 
 unique_ptr<TableRef> ParquetScanReplacement(ClientContext &context, const string &table_name,
@@ -714,7 +716,7 @@ void ParquetExtension::Load(DuckDB &db) {
 	function.copy_to_sink = ParquetWriteSink;
 	function.copy_to_combine = ParquetWriteCombine;
 	function.copy_to_finalize = ParquetWriteFinalize;
-	function.parallel = ParquetWriteIsParallel;
+	function.execution_mode = ParquetWriteExecutionMode;
 	function.copy_from_bind = ParquetScanFunction::ParquetReadBind;
 	function.copy_from_function = scan_fun.functions[0];
 

package/src/duckdb/src/catalog/catalog_entry/table_catalog_entry.cpp

@@ -51,16 +51,27 @@ vector<LogicalType> TableCatalogEntry::GetTypes() {
 	return types;
 }
 
+CreateTableInfo TableCatalogEntry::GetTableInfoForSerialization() const {
+	CreateTableInfo result;
+	result.catalog = catalog.GetName();
+	result.schema = schema.name;
+	result.table = name;
+	result.columns = columns.Copy();
+	result.constraints.reserve(constraints.size());
+	std::for_each(constraints.begin(), constraints.end(),
+	              [&result](const unique_ptr<Constraint> &c) { result.constraints.emplace_back(c->Copy()); });
+	return result;
+}
+
 void TableCatalogEntry::Serialize(Serializer &serializer) const {
 	D_ASSERT(!internal);
-
+	const auto info = GetTableInfoForSerialization();
 	FieldWriter writer(serializer);
-	auto catalog_name = catalog.GetName();
-	writer.WriteString(catalog_name);
-	writer.WriteString(schema.name);
-	writer.WriteString(name);
-	columns.Serialize(writer);
-	writer.WriteSerializableList(constraints);
+	writer.WriteString(info.catalog);
+	writer.WriteString(info.schema);
+	writer.WriteString(info.table);
+	info.columns.Serialize(writer);
+	writer.WriteSerializableList(info.constraints);
 	writer.Finalize();
 }
 

package/src/duckdb/src/catalog/default/default_functions.cpp

@@ -79,6 +79,8 @@ static DefaultMacro internal_macros[] = {
 	{"pg_catalog", "pg_ts_template_is_visible", {"template_oid", nullptr}, "true"},
 	{"pg_catalog", "pg_type_is_visible", {"type_oid", nullptr}, "true"},
 
+	{"pg_catalog", "pg_size_pretty", {"bytes", nullptr}, "format_bytes(bytes)"},
+
 	{DEFAULT_SCHEMA, "round_even", {"x", "n", nullptr}, "CASE ((abs(x) * power(10, n+1)) % 10) WHEN 5 THEN round(x/2, n) * 2 ELSE round(x, n) END"},
 	{DEFAULT_SCHEMA, "roundbankers", {"x", "n", nullptr}, "round_even(x, n)"},
 	{DEFAULT_SCHEMA, "nullif", {"a", "b", nullptr}, "CASE WHEN a=b THEN NULL ELSE a END"},

package/src/duckdb/src/common/arrow/arrow_appender.cpp

@@ -320,7 +320,7 @@ struct ArrowUUIDConverter {
 template <class SRC = string_t, class OP = ArrowVarcharConverter>
 struct ArrowVarcharData {
 	static void Initialize(ArrowAppendData &result, const LogicalType &type, idx_t capacity) {
-		result.main_buffer.reserve((capacity + 1) * sizeof(uint32_t));
+		result.main_buffer.reserve((capacity + 1) * sizeof(uint64_t));
 		result.aux_buffer.reserve(capacity);
 	}
 
@@ -334,9 +334,9 @@ struct ArrowVarcharData {
 		auto validity_data = (uint8_t *)append_data.validity.data();
 
 		// resize the offset buffer - the offset buffer holds the offsets into the child array
-		append_data.main_buffer.resize(append_data.main_buffer.size() + sizeof(uint32_t) * (size + 1));
+		append_data.main_buffer.resize(append_data.main_buffer.size() + sizeof(uint64_t) * (size + 1));
 		auto data = (SRC *)format.data;
-		auto offset_data = (uint32_t *)append_data.main_buffer.data();
+		auto offset_data = (uint64_t *)append_data.main_buffer.data();
 		if (append_data.row_count == 0) {
 			// first entry
 			offset_data[0] = 0;

package/src/duckdb/src/common/arrow/arrow_converter.cpp

@@ -117,7 +117,7 @@ void SetArrowFormat(DuckDBArrowSchemaHolder &root_holder, ArrowSchema &child, co
 		break;
 	case LogicalTypeId::UUID:
 	case LogicalTypeId::VARCHAR:
-		child.format = "u";
+		child.format = "U";
 		break;
 	case LogicalTypeId::DATE:
 		child.format = "tdD";
@@ -171,7 +171,7 @@ void SetArrowFormat(DuckDBArrowSchemaHolder &root_holder, ArrowSchema &child, co
 	}
 	case LogicalTypeId::BLOB:
 	case LogicalTypeId::BIT: {
-		child.format = "z";
+		child.format = "Z";
 		break;
 	}
 	case LogicalTypeId::LIST: {