duckdb 0.7.1-dev90.0 → 0.7.2-dev0.0

package/src/duckdb/extension/json/json_scan.cpp

@@ -1,9 +1,9 @@
 #include "json_scan.hpp"
 
 #include "duckdb/main/database.hpp"
+#include "duckdb/main/extension_helper.hpp"
 #include "duckdb/parallel/task_scheduler.hpp"
 #include "duckdb/storage/buffer_manager.hpp"
-#include "duckdb/main/extension_helper.hpp"
 
 namespace duckdb {
 
@@ -20,8 +20,9 @@ unique_ptr<FunctionData> JSONScanData::Bind(ClientContext &context, TableFunctio
 	auto &options = result->options;
 
 	auto &info = (JSONScanInfo &)*input.info;
-	options.format = info.format;
 	result->type = info.type;
+	options.format = info.format;
+	result->record_type = info.record_type;
 	result->auto_detect = info.auto_detect;
 
 	vector<string> patterns;
@@ -40,16 +41,16 @@ unique_ptr<FunctionData> JSONScanData::Bind(ClientContext &context, TableFunctio
 			result->ignore_errors = BooleanValue::Get(kv.second);
 		} else if (loption == "maximum_object_size") {
 			result->maximum_object_size = MaxValue<idx_t>(UIntegerValue::Get(kv.second), result->maximum_object_size);
-		} else if (loption == "format") {
+		} else if (loption == "lines") {
 			auto format = StringUtil::Lower(StringValue::Get(kv.second));
 			if (format == "auto") {
 				options.format = JSONFormat::AUTO_DETECT;
-			} else if (format == "unstructured") {
+			} else if (format == "false") {
 				options.format = JSONFormat::UNSTRUCTURED;
-			} else if (format == "newline_delimited") {
+			} else if (format == "true") {
 				options.format = JSONFormat::NEWLINE_DELIMITED;
 			} else {
-				throw BinderException("format must be one of ['auto', 'unstructured', 'newline_delimited']");
+				throw BinderException("\"lines\" must be one of ['auto', 'true', 'false']");
 			}
 		} else if (loption == "compression") {
 			auto compression = StringUtil::Lower(StringValue::Get(kv.second));
@@ -74,10 +75,7 @@ void JSONScanData::InitializeFilePaths(ClientContext &context, const vector<stri
                                        vector<string> &file_paths) {
 	auto &fs = FileSystem::GetFileSystem(context);
 	for (auto &file_pattern : patterns) {
-		auto found_files = fs.Glob(file_pattern, context);
-		if (found_files.empty()) {
-			throw FileSystem::MissingFileException(file_pattern, context);
-		}
+		auto found_files = fs.GlobFiles(file_pattern, context);
 		file_paths.insert(file_paths.end(), found_files.begin(), found_files.end());
 	}
 }
@@ -98,6 +96,27 @@ void JSONScanData::InitializeFormats() {
 	if (!timestamp_format.empty()) {
 		date_format_map.AddFormat(LogicalTypeId::TIMESTAMP, timestamp_format);
 	}
+
+	if (auto_detect) {
+		static const unordered_map<LogicalTypeId, vector<const char *>, LogicalTypeIdHash> FORMAT_TEMPLATES = {
+		    {LogicalTypeId::DATE, {"%m-%d-%Y", "%m-%d-%y", "%d-%m-%Y", "%d-%m-%y", "%Y-%m-%d", "%y-%m-%d"}},
+		    {LogicalTypeId::TIMESTAMP,
+		     {"%Y-%m-%d %H:%M:%S.%f", "%m-%d-%Y %I:%M:%S %p", "%m-%d-%y %I:%M:%S %p", "%d-%m-%Y %H:%M:%S",
+		      "%d-%m-%y %H:%M:%S", "%Y-%m-%d %H:%M:%S", "%y-%m-%d %H:%M:%S", "%Y-%m-%dT%H:%M:%SZ"}},
+		};
+
+		// Populate possible date/timestamp formats, assume this is consistent across columns
+		for (auto &kv : FORMAT_TEMPLATES) {
+			const auto &type = kv.first;
+			if (date_format_map.HasFormats(type)) {
+				continue; // Already populated
+			}
+			const auto &format_strings = kv.second;
+			for (auto &format_string : format_strings) {
+				date_format_map.AddFormat(type, format_string);
+			}
+		}
+	}
 }
 
 void JSONScanData::Serialize(FieldWriter &writer) {
@@ -112,9 +131,17 @@ void JSONScanData::Serialize(FieldWriter &writer) {
 	writer.WriteList<string>(names);
 	writer.WriteList<idx_t>(valid_cols);
 	writer.WriteField<idx_t>(max_depth);
-	writer.WriteField<bool>(objects);
-	writer.WriteString(date_format);
-	writer.WriteString(timestamp_format);
+	writer.WriteField<JSONRecordType>(record_type);
+	if (!date_format.empty()) {
+		writer.WriteString(date_format);
+	} else {
+		writer.WriteString(date_format_map.GetFormat(LogicalTypeId::DATE).format_specifier);
+	}
+	if (!timestamp_format.empty()) {
+		writer.WriteString(timestamp_format);
+	} else {
+		writer.WriteString(date_format_map.GetFormat(LogicalTypeId::TIMESTAMP).format_specifier);
+	}
 }
 
 void JSONScanData::Deserialize(FieldReader &reader) {
@@ -129,9 +156,12 @@ void JSONScanData::Deserialize(FieldReader &reader) {
 	names = reader.ReadRequiredList<string>();
 	valid_cols = reader.ReadRequiredList<idx_t>();
 	max_depth = reader.ReadRequired<idx_t>();
-	objects = reader.ReadRequired<bool>();
+	record_type = reader.ReadRequired<JSONRecordType>();
 	date_format = reader.ReadRequired<string>();
 	timestamp_format = reader.ReadRequired<string>();
+
+	InitializeFormats();
+	transform_options.date_format_map = &date_format_map;
 }
 
 JSONScanGlobalState::JSONScanGlobalState(ClientContext &context, JSONScanData &bind_data_p)
@@ -150,11 +180,11 @@ JSONScanGlobalState::JSONScanGlobalState(ClientContext &context, JSONScanData &b
 }
 
 JSONScanLocalState::JSONScanLocalState(ClientContext &context, JSONScanGlobalState &gstate)
-    : batch_index(DConstants::INVALID_INDEX), bind_data(gstate.bind_data),
+    : scan_count(0), array_idx(0), array_offset(0), batch_index(DConstants::INVALID_INDEX), bind_data(gstate.bind_data),
       json_allocator(BufferAllocator::Get(context)), current_reader(nullptr), current_buffer_handle(nullptr),
-      buffer_size(0), buffer_offset(0), prev_buffer_remainder(0) {
+      is_last(false), buffer_size(0), buffer_offset(0), prev_buffer_remainder(0) {
 
-	// Buffer to reconstruct JSON objects when they cross a buffer boundary
+	// Buffer to reconstruct JSON values when they cross a buffer boundary
 	reconstruct_buffer = gstate.allocator.Allocate(gstate.bind_data.maximum_object_size + YYJSON_PADDING_SIZE);
 
 	// This is needed for JSONFormat::UNSTRUCTURED, to make use of YYJSON_READ_INSITU
@@ -174,11 +204,6 @@ unique_ptr<GlobalTableFunctionState> JSONGlobalTableFunctionState::Init(ClientCo
 	// Perform projection pushdown
 	if (bind_data.type == JSONScanType::READ_JSON) {
 		D_ASSERT(input.column_ids.size() <= bind_data.names.size()); // Can't project to have more columns
-		if (bind_data.auto_detect && input.column_ids.size() < bind_data.names.size()) {
-			// If we are auto-detecting, but don't need all columns present in the file,
-			// then we don't need to throw an error if we encounter an unseen column
-			bind_data.transform_options.error_unknown_key = false;
-		}
 		vector<string> names;
 		names.reserve(input.column_ids.size());
 		for (idx_t i = 0; i < input.column_ids.size(); i++) {
@@ -189,13 +214,37 @@ unique_ptr<GlobalTableFunctionState> JSONGlobalTableFunctionState::Init(ClientCo
 			names.push_back(std::move(bind_data.names[id]));
 			bind_data.valid_cols.push_back(i);
 		}
+		if (names.size() < bind_data.names.size()) {
+			// If we are auto-detecting, but don't need all columns present in the file,
+			// then we don't need to throw an error if we encounter an unseen column
+			bind_data.transform_options.error_unknown_key = false;
+		}
 		bind_data.names = std::move(names);
 	}
 	return result;
 }
 
 idx_t JSONGlobalTableFunctionState::MaxThreads() const {
-	return state.system_threads;
+	auto &bind_data = state.bind_data;
+
+	auto num_files = bind_data.file_paths.size();
+	idx_t readers_per_file;
+	if (bind_data.options.format == JSONFormat::UNSTRUCTURED) {
+		// Unstructured necessitates single thread
+		readers_per_file = 1;
+	} else if (!state.json_readers.empty() && state.json_readers[0]->IsOpen()) {
+		auto &reader = *state.json_readers[0];
+		const auto &options = reader.GetOptions();
+		if (options.format == JSONFormat::UNSTRUCTURED || options.compression != FileCompressionType::UNCOMPRESSED) {
+			// Auto-detected unstructured - same story, compression also really limits parallelism
+			readers_per_file = 1;
+		} else {
+			return state.system_threads;
+		}
+	} else {
+		return state.system_threads;
+	}
+	return num_files * readers_per_file;
 }
 
 JSONLocalTableFunctionState::JSONLocalTableFunctionState(ClientContext &context, JSONScanGlobalState &gstate)
@@ -231,6 +280,12 @@ static inline void SkipWhitespace(const char *buffer_ptr, idx_t &buffer_offset,
 idx_t JSONScanLocalState::ReadNext(JSONScanGlobalState &gstate) {
 	json_allocator.Reset();
 
+	if ((gstate.bind_data.record_type == JSONRecordType::ARRAY_OF_RECORDS ||
+	     gstate.bind_data.record_type == JSONRecordType::ARRAY_OF_JSON) &&
+	    array_idx < scan_count) {
+		return GetObjectsFromArray(gstate);
+	}
+
 	idx_t count = 0;
 	if (buffer_offset == buffer_size) {
 		if (!ReadNextBuffer(gstate)) {
@@ -254,10 +309,18 @@ idx_t JSONScanLocalState::ReadNext(JSONScanGlobalState &gstate) {
 	default:
 		throw InternalException("Unknown JSON format");
 	}
+	scan_count = count;
 
 	// Skip over any remaining whitespace for the next scan
 	SkipWhitespace(buffer_ptr, buffer_offset, buffer_size);
 
+	if (gstate.bind_data.record_type == JSONRecordType::ARRAY_OF_RECORDS ||
+	    gstate.bind_data.record_type == JSONRecordType::ARRAY_OF_JSON) {
+		array_idx = 0;
+		array_offset = 0;
+		return GetObjectsFromArray(gstate);
+	}
+
 	return count;
 }
 
@@ -332,10 +395,48 @@ yyjson_val *JSONScanLocalState::ParseLine(char *line_start, idx_t line_size, idx
 	}
 }
 
+idx_t JSONScanLocalState::GetObjectsFromArray(JSONScanGlobalState &gstate) {
+	idx_t arr_count = 0;
+
+	size_t idx, max;
+	yyjson_val *val;
+	for (; array_idx < scan_count; array_idx++, array_offset = 0) {
+		auto &value = values[array_idx];
+		if (!value) {
+			continue;
+		}
+		if (unsafe_yyjson_is_arr(value)) {
+			yyjson_arr_foreach(value, idx, max, val) {
+				if (idx < array_offset) {
+					continue;
+				}
+				array_values[arr_count++] = val;
+				if (arr_count == STANDARD_VECTOR_SIZE) {
+					break;
+				}
+			}
+			array_offset = idx + 1;
+			if (arr_count == STANDARD_VECTOR_SIZE) {
+				break;
+			}
+		} else if (!gstate.bind_data.ignore_errors) {
+			ThrowTransformError(
+			    array_idx,
+			    StringUtil::Format("Expected JSON ARRAY but got %s: %s\nTry setting json_format to 'records'",
+			                       JSONCommon::ValTypeToString(value), JSONCommon::ValToString(value, 50)));
+		}
+	}
+	return arr_count;
+}
+
 bool JSONScanLocalState::ReadNextBuffer(JSONScanGlobalState &gstate) {
 	if (current_reader) {
 		D_ASSERT(current_buffer_handle);
 		current_reader->SetBufferLineOrObjectCount(current_buffer_handle->buffer_index, lines_or_objects_in_buffer);
+		if (is_last && gstate.bind_data.type != JSONScanType::SAMPLE) {
+			// Close files that are done if we're not sampling
+			current_reader->CloseJSONFile();
+		}
 	}
 
 	AllocatedData buffer;
@@ -396,7 +497,9 @@ bool JSONScanLocalState::ReadNextBuffer(JSONScanGlobalState &gstate) {
 		// Unopened file
 		current_reader->OpenJSONFile();
 		batch_index = gstate.batch_index++;
-		if (options.format == JSONFormat::UNSTRUCTURED) {
+		if (options.format == JSONFormat::UNSTRUCTURED || (options.format == JSONFormat::NEWLINE_DELIMITED &&
+		                                                   options.compression != FileCompressionType::UNCOMPRESSED &&
+		                                                   gstate.file_index < gstate.json_readers.size())) {
 			gstate.file_index++; // UNSTRUCTURED necessitates single-threaded read
 		}
 		if (options.format != JSONFormat::AUTO_DETECT) {
@@ -450,9 +553,6 @@ bool JSONScanLocalState::ReadNextBuffer(JSONScanGlobalState &gstate) {
 	auto json_buffer_handle = make_unique<JSONBufferHandle>(buffer_index, readers, std::move(buffer), buffer_size);
 	current_buffer_handle = json_buffer_handle.get();
 	current_reader->InsertBuffer(buffer_index, std::move(json_buffer_handle));
-	if (!current_reader->GetFileHandle().PlainFileSource() && gstate.bind_data.type == JSONScanType::SAMPLE) {
-		// TODO: store buffer
-	}
 
 	buffer_offset = 0;
 	prev_buffer_remainder = 0;
@@ -508,16 +608,18 @@ void JSONScanLocalState::ReadNextBufferSeek(JSONScanGlobalState &gstate, idx_t &
 }
 
 void JSONScanLocalState::ReadNextBufferNoSeek(JSONScanGlobalState &gstate, idx_t &buffer_index) {
-	auto &file_handle = current_reader->GetFileHandle();
-
 	idx_t request_size = gstate.buffer_capacity - prev_buffer_remainder - YYJSON_PADDING_SIZE;
 	idx_t read_size;
 	{
 		lock_guard<mutex> reader_guard(current_reader->lock);
 		buffer_index = current_reader->GetBufferIndex();
 
-		read_size = file_handle.Read(buffer_ptr + prev_buffer_remainder, request_size,
-		                             gstate.bind_data.type == JSONScanType::SAMPLE);
+		if (current_reader->IsOpen()) {
+			read_size = current_reader->GetFileHandle().Read(buffer_ptr + prev_buffer_remainder, request_size,
+			                                                 gstate.bind_data.type == JSONScanType::SAMPLE);
+		} else {
+			read_size = 0;
+		}
 		is_last = read_size < request_size;
 
 		if (!gstate.bind_data.ignore_errors && read_size == 0 && prev_buffer_remainder != 0) {
@@ -579,10 +681,15 @@ void JSONScanLocalState::ReconstructFirstObject(JSONScanGlobalState &gstate) {
 		current_reader->RemoveBuffer(current_buffer_handle->buffer_index - 1);
 	}
 
-	objects[0] = ParseLine((char *)reconstruct_ptr, line_size, line_size, lines[0]);
+	values[0] = ParseLine((char *)reconstruct_ptr, line_size, line_size, lines[0]);
 }
 
 void JSONScanLocalState::ReadUnstructured(idx_t &count) {
+	// yyjson does not always return YYJSON_READ_ERROR_UNEXPECTED_END properly
+	// if a different error code happens within the last 50 bytes
+	// we assume it should be YYJSON_READ_ERROR_UNEXPECTED_END instead
+	static constexpr idx_t END_BOUND = 50;
+
 	const auto max_obj_size = reconstruct_buffer.GetSize();
 	yyjson_read_err error;
 	for (; count < STANDARD_VECTOR_SIZE; count++) {
@@ -608,8 +715,7 @@ void JSONScanLocalState::ReadUnstructured(idx_t &count) {
 		} else if (error.pos > max_obj_size) {
 			current_reader->ThrowParseError(current_buffer_handle->buffer_index, lines_or_objects_in_buffer, error,
 			                                "Try increasing \"maximum_object_size\".");
-
-		} else if (error.code == YYJSON_READ_ERROR_UNEXPECTED_END && !is_last) {
+		} else if (!is_last && (error.code == YYJSON_READ_ERROR_UNEXPECTED_END || remaining - error.pos < END_BOUND)) {
 			// Copy remaining to reconstruct_buffer
 			const auto reconstruct_ptr = reconstruct_buffer.get();
 			memcpy(reconstruct_ptr, obj_copy_start, remaining);
@@ -619,7 +725,7 @@ void JSONScanLocalState::ReadUnstructured(idx_t &count) {
 		} else {
 			current_reader->ThrowParseError(current_buffer_handle->buffer_index, lines_or_objects_in_buffer, error);
 		}
-		objects[count] = read_doc->root;
+		values[count] = read_doc->root;
 	}
 }
 
@@ -645,7 +751,7 @@ void JSONScanLocalState::ReadNewlineDelimited(idx_t &count) {
 		}
 		idx_t line_size = line_end - line_start;
 
-		objects[count] = ParseLine((char *)line_start, line_size, remaining, lines[count]);
+		values[count] = ParseLine((char *)line_start, line_size, remaining, lines[count]);
 
 		buffer_offset += line_size;
 		SkipWhitespace(buffer_ptr, buffer_offset, buffer_size);
@@ -656,11 +762,11 @@ yyjson_alc *JSONScanLocalState::GetAllocator() {
 	return json_allocator.GetYYJSONAllocator();
 }
 
-void JSONScanLocalState::ThrowTransformError(idx_t count, idx_t object_index, const string &error_message) {
+void JSONScanLocalState::ThrowTransformError(idx_t object_index, const string &error_message) {
 	D_ASSERT(current_reader);
 	D_ASSERT(current_buffer_handle);
 	D_ASSERT(object_index != DConstants::INVALID_INDEX);
-	auto line_or_object_in_buffer = lines_or_objects_in_buffer - count + object_index;
+	auto line_or_object_in_buffer = lines_or_objects_in_buffer - scan_count + object_index;
 	current_reader->ThrowTransformError(current_buffer_handle->buffer_index, line_or_object_in_buffer, error_message);
 }
 

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

@@ -589,6 +589,7 @@ void StringColumnReader::PrepareDeltaLengthByteArray(ResizeableBuffer &buffer) {
 	}
 	auto length_data = (uint32_t *)length_buffer->ptr;
 	byte_array_data = make_unique<Vector>(LogicalType::VARCHAR, value_count);
+	byte_array_count = value_count;
 	auto string_data = FlatVector::GetData<string_t>(*byte_array_data);
 	for (idx_t i = 0; i < value_count; i++) {
 		auto str_len = length_data[i];
@@ -615,6 +616,7 @@ void StringColumnReader::PrepareDeltaByteArray(ResizeableBuffer &buffer) {
 	auto prefix_data = (uint32_t *)prefix_buffer->ptr;
 	auto suffix_data = (uint32_t *)suffix_buffer->ptr;
 	byte_array_data = make_unique<Vector>(LogicalType::VARCHAR, prefix_count);
+	byte_array_count = prefix_count;
 	auto string_data = FlatVector::GetData<string_t>(*byte_array_data);
 	for (idx_t i = 0; i < prefix_count; i++) {
 		auto str_len = prefix_data[i] + suffix_data[i];
@@ -646,6 +648,11 @@ void StringColumnReader::DeltaByteArray(uint8_t *defines, idx_t num_values, parq
 			continue;
 		}
 		if (filter[row_idx + result_offset]) {
+			if (delta_offset >= byte_array_count) {
+				throw IOException("DELTA_BYTE_ARRAY - length mismatch between values and byte array lengths (attempted "
+				                  "read of %d from %d entries) - corrupt file?",
+				                  delta_offset + 1, byte_array_count);
+			}
 			result_ptr[row_idx + result_offset] = string_data[delta_offset++];
 		} else {
 			delta_offset++;

package/src/duckdb/extension/parquet/include/column_reader.hpp

@@ -131,6 +131,7 @@ protected:
 	ParquetReader &reader;
 	LogicalType type;
 	unique_ptr<Vector> byte_array_data;
+	idx_t byte_array_count = 0;
 
 	idx_t pending_skips = 0;
 

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

@@ -221,10 +221,7 @@ public:
 		}
 
 		FileSystem &fs = FileSystem::GetFileSystem(context);
-		auto files = fs.Glob(info.file_path, context);
-		if (files.empty()) {
-			throw FileSystem::MissingFileException(info.file_path, context);
-		}
+		auto files = fs.GlobFiles(info.file_path, context);
 
 		// The most likely path (Parquet read without union by name option)
 		if (!parquet_options.union_by_name) {
@@ -362,12 +359,7 @@ public:
 	}
 
 	static vector<string> ParquetGlob(FileSystem &fs, const string &glob, ClientContext &context) {
-		auto files = fs.Glob(glob, FileSystem::GetFileOpener(context));
-
-		if (files.empty()) {
-			throw FileSystem::MissingFileException(glob, context);
-		}
-		return files;
+		return fs.GlobFiles(glob, context);
 	}
 
 	static unique_ptr<FunctionData> ParquetScanBind(ClientContext &context, TableFunctionBindInput &input,

package/src/duckdb/src/catalog/catalog.cpp

@@ -2,6 +2,7 @@
 
 #include "duckdb/catalog/catalog_search_path.hpp"
 #include "duckdb/catalog/catalog_entry/list.hpp"
+#include "duckdb/catalog/catalog_entry/table_catalog_entry.hpp"
 #include "duckdb/catalog/catalog_set.hpp"
 #include "duckdb/catalog/default/default_schemas.hpp"
 #include "duckdb/catalog/catalog_entry/type_catalog_entry.hpp"
@@ -26,7 +27,7 @@
 #include "duckdb/planner/parsed_data/bound_create_table_info.hpp"
 #include "duckdb/planner/binder.hpp"
 #include "duckdb/catalog/default/default_types.hpp"
-#include "duckdb/main/extension_functions.hpp"
+#include "duckdb/main/extension_entries.hpp"
 #include "duckdb/main/connection.hpp"
 #include "duckdb/main/attached_database.hpp"
 #include "duckdb/main/database_manager.hpp"
@@ -251,6 +252,20 @@ CatalogEntry *Catalog::CreateCollation(CatalogTransaction transaction, SchemaCat
 	return schema->CreateCollation(transaction, info);
 }
 
+//===--------------------------------------------------------------------===//
+// Index
+//===--------------------------------------------------------------------===//
+CatalogEntry *Catalog::CreateIndex(CatalogTransaction transaction, CreateIndexInfo *info) {
+	auto &context = transaction.GetContext();
+	return CreateIndex(context, info);
+}
+
+CatalogEntry *Catalog::CreateIndex(ClientContext &context, CreateIndexInfo *info) {
+	auto schema = GetSchema(context, info->schema);
+	auto table = GetEntry<TableCatalogEntry>(context, schema->name, info->table->table_name);
+	return schema->CreateIndex(context, info, table);
+}
+
 //===--------------------------------------------------------------------===//
 // Lookup Structures
 //===--------------------------------------------------------------------===//
@@ -317,17 +332,26 @@ SimilarCatalogEntry Catalog::SimilarEntryInSchemas(ClientContext &context, const
 	return result;
 }
 
-string FindExtension(const string &function_name) {
-	auto size = sizeof(EXTENSION_FUNCTIONS) / sizeof(ExtensionFunction);
-	auto it = std::lower_bound(
-	    EXTENSION_FUNCTIONS, EXTENSION_FUNCTIONS + size, function_name,
-	    [](const ExtensionFunction &element, const string &value) { return element.function < value; });
-	if (it != EXTENSION_FUNCTIONS + size && it->function == function_name) {
+string FindExtensionGeneric(const string &name, const ExtensionEntry entries[], idx_t size) {
+	auto lcase = StringUtil::Lower(name);
+	auto it = std::lower_bound(entries, entries + size, lcase,
+	                           [](const ExtensionEntry &element, const string &value) { return element.name < value; });
+	if (it != entries + size && it->name == lcase) {
 		return it->extension;
 	}
 	return "";
 }
 
+string FindExtensionForFunction(const string &name) {
+	idx_t size = sizeof(EXTENSION_FUNCTIONS) / sizeof(ExtensionEntry);
+	return FindExtensionGeneric(name, EXTENSION_FUNCTIONS, size);
+}
+
+string FindExtensionForSetting(const string &name) {
+	idx_t size = sizeof(EXTENSION_SETTINGS) / sizeof(ExtensionEntry);
+	return FindExtensionGeneric(name, EXTENSION_SETTINGS, size);
+}
+
 vector<CatalogSearchEntry> GetCatalogEntries(ClientContext &context, const string &catalog, const string &schema) {
 	vector<CatalogSearchEntry> entries;
 	auto &search_path = *context.client_data->catalog_search_path;
@@ -392,6 +416,26 @@ void FindMinimalQualification(ClientContext &context, const string &catalog_name
 	qualify_schema = true;
 }
 
+CatalogException Catalog::UnrecognizedConfigurationError(ClientContext &context, const string &name) {
+	// check if the setting exists in any extensions
+	auto extension_name = FindExtensionForSetting(name);
+	if (!extension_name.empty()) {
+		return CatalogException(
+		    "Setting with name \"%s\" is not in the catalog, but it exists in the %s extension.\n\nTo "
+		    "install and load the extension, run:\nINSTALL %s;\nLOAD %s;",
+		    name, extension_name, extension_name, extension_name);
+	}
+	// the setting is not in an extension
+	// get a list of all options
+	vector<string> potential_names = DBConfig::GetOptionNames();
+	for (auto &entry : DBConfig::GetConfig(context).extension_parameters) {
+		potential_names.push_back(entry.first);
+	}
+
+	throw CatalogException("unrecognized configuration parameter \"%s\"\n%s", name,
+	                       StringUtil::CandidatesErrorMessage(potential_names, name, "Did you mean"));
+}
+
 CatalogException Catalog::CreateMissingEntryException(ClientContext &context, const string &entry_name,
                                                       CatalogType type,
                                                       const unordered_set<SchemaCatalogEntry *> &schemas,
@@ -408,13 +452,18 @@ CatalogException Catalog::CreateMissingEntryException(ClientContext &context, co
 			unseen_schemas.insert(current_schema);
 		}
 	}
-	auto unseen_entry = SimilarEntryInSchemas(context, entry_name, type, unseen_schemas);
-	auto extension_name = FindExtension(entry_name);
-	if (!extension_name.empty()) {
-		return CatalogException("Function with name %s is not on the catalog, but it exists in the %s extension. To "
-		                        "Install and Load the extension, run: INSTALL %s; LOAD %s;",
-		                        entry_name, extension_name, extension_name, extension_name);
+	// check if the entry exists in any extension
+	if (type == CatalogType::TABLE_FUNCTION_ENTRY || type == CatalogType::SCALAR_FUNCTION_ENTRY ||
+	    type == CatalogType::AGGREGATE_FUNCTION_ENTRY) {
+		auto extension_name = FindExtensionForFunction(entry_name);
+		if (!extension_name.empty()) {
+			return CatalogException(
+			    "Function with name \"%s\" is not in the catalog, but it exists in the %s extension.\n\nTo "
+			    "install and load the extension, run:\nINSTALL %s;\nLOAD %s;",
+			    entry_name, extension_name, extension_name, extension_name);
+		}
 	}
+	auto unseen_entry = SimilarEntryInSchemas(context, entry_name, type, unseen_schemas);
 	string did_you_mean;
 	if (unseen_entry.Found() && unseen_entry.distance < entry.distance) {
 		// the closest matching entry requires qualification as it is not in the default search path

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

@@ -19,10 +19,11 @@ string IndexCatalogEntry::ToSQL() {
 	return sql;
 }
 
-void IndexCatalogEntry::Serialize(duckdb::MetaBlockWriter &serializer) {
-	// Here we serialize the index metadata in the following order:
-	// schema name, table name, index name, sql, index type, index constraint type, expression list.
-	// column_ids, unbound_expression
+void IndexCatalogEntry::Serialize(Serializer &serializer) {
+	// here we serialize the index metadata in the following order:
+	// schema name, table name, index name, sql, index type, index constraint type, expression list, parsed expressions,
+	// column IDs
+
 	FieldWriter writer(serializer);
 	writer.WriteString(GetSchemaName());
 	writer.WriteString(GetTableName());
@@ -37,9 +38,9 @@ void IndexCatalogEntry::Serialize(duckdb::MetaBlockWriter &serializer) {
 }
 
 unique_ptr<CreateIndexInfo> IndexCatalogEntry::Deserialize(Deserializer &source, ClientContext &context) {
-	// Here we deserialize the index metadata in the following order:
-	// root block, root offset, schema name, table name, index name, sql, index type, index constraint type, expression
-	// list.
+	// here we deserialize the index metadata in the following order:
+	// schema name, table schema name, table name, index name, sql, index type, index constraint type, expression list,
+	// parsed expression list, column IDs
 
 	auto create_index_info = make_unique<CreateIndexInfo>();
 

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

@@ -24,7 +24,7 @@ SimilarCatalogEntry SchemaCatalogEntry::GetSimilarEntry(CatalogTransaction trans
                                                         const string &name) {
 	SimilarCatalogEntry result;
 	Scan(transaction.GetContext(), type, [&](CatalogEntry *entry) {
-		auto ldist = StringUtil::LevenshteinDistance(entry->name, name);
+		auto ldist = StringUtil::SimilarityScore(entry->name, name);
 		if (ldist < result.distance) {
 			result.distance = ldist;
 			result.name = entry->name;

package/src/duckdb/src/catalog/catalog_set.cpp

@@ -460,7 +460,7 @@ SimilarCatalogEntry CatalogSet::SimilarEntry(CatalogTransaction transaction, con
 	for (auto &kv : mapping) {
 		auto mapping_value = GetMapping(transaction, kv.first);
 		if (mapping_value && !mapping_value->deleted) {
-			auto ldist = StringUtil::LevenshteinDistance(kv.first, name);
+			auto ldist = StringUtil::SimilarityScore(kv.first, name);
 			if (ldist < result.distance) {
 				result.distance = ldist;
 				result.name = kv.first;

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

@@ -93,6 +93,7 @@ static DefaultMacro internal_macros[] = {
 	{DEFAULT_SCHEMA, "fdiv", {"x", "y", nullptr}, "floor(x/y)"},
 	{DEFAULT_SCHEMA, "fmod", {"x", "y", nullptr}, "(x-y*floor(x/y))"},
 	{DEFAULT_SCHEMA, "count_if", {"l", nullptr}, "sum(if(l, 1, 0))"},
+	{DEFAULT_SCHEMA, "split_part", {"string", "delimiter", "position", nullptr}, "coalesce(string_split(string, delimiter)[position],'')"},
 
 	// algebraic list aggregates
 	{DEFAULT_SCHEMA, "list_avg", {"l", nullptr}, "list_aggr(l, 'avg')"},

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

@@ -48,7 +48,7 @@ static DefaultView internal_views[] = {
     {"pg_catalog", "pg_views", "SELECT schema_name schemaname, view_name viewname, 'duckdb' viewowner, sql definition FROM duckdb_views()"},
     {"information_schema", "columns", "SELECT database_name table_catalog, schema_name table_schema, table_name, column_name, column_index ordinal_position, column_default, CASE WHEN is_nullable THEN 'YES' ELSE 'NO' END is_nullable, data_type, character_maximum_length, NULL character_octet_length, numeric_precision, numeric_precision_radix, numeric_scale, NULL datetime_precision, NULL interval_type, NULL interval_precision, NULL character_set_catalog, NULL character_set_schema, NULL character_set_name, NULL collation_catalog, NULL collation_schema, NULL collation_name, NULL domain_catalog, NULL domain_schema, NULL domain_name, NULL udt_catalog, NULL udt_schema, NULL udt_name, NULL scope_catalog, NULL scope_schema, NULL scope_name, NULL maximum_cardinality, NULL dtd_identifier, NULL is_self_referencing, NULL is_identity, NULL identity_generation, NULL identity_start, NULL identity_increment, NULL identity_maximum, NULL identity_minimum, NULL identity_cycle, NULL is_generated, NULL generation_expression, NULL is_updatable FROM duckdb_columns;"},
     {"information_schema", "schemata", "SELECT database_name catalog_name, schema_name, 'duckdb' schema_owner, NULL default_character_set_catalog, NULL default_character_set_schema, NULL default_character_set_name, sql sql_path FROM duckdb_schemas()"},
-    {"information_schema", "tables", "SELECT database_name table_catalog, schema_name table_schema, table_name, CASE WHEN temporary THEN 'LOCAL TEMPORARY' ELSE 'BASE TABLE' END table_type, NULL self_referencing_column_name, NULL reference_generation, NULL user_defined_type_catalog, NULL user_defined_type_schema, NULL user_defined_type_name, 'YES' is_insertable_into, 'NO' is_typed, CASE WHEN temporary THEN 'PRESERVE' ELSE NULL END commit_action FROM duckdb_tables() UNION ALL SELECT NULL table_catalog, schema_name table_schema, view_name table_name, 'VIEW' table_type, NULL self_referencing_column_name, NULL reference_generation, NULL user_defined_type_catalog, NULL user_defined_type_schema, NULL user_defined_type_name, 'NO' is_insertable_into, 'NO' is_typed, NULL commit_action FROM duckdb_views;"},
+    {"information_schema", "tables", "SELECT database_name table_catalog, schema_name table_schema, table_name, CASE WHEN temporary THEN 'LOCAL TEMPORARY' ELSE 'BASE TABLE' END table_type, NULL self_referencing_column_name, NULL reference_generation, NULL user_defined_type_catalog, NULL user_defined_type_schema, NULL user_defined_type_name, 'YES' is_insertable_into, 'NO' is_typed, CASE WHEN temporary THEN 'PRESERVE' ELSE NULL END commit_action FROM duckdb_tables() UNION ALL SELECT database_name table_catalog, schema_name table_schema, view_name table_name, 'VIEW' table_type, NULL self_referencing_column_name, NULL reference_generation, NULL user_defined_type_catalog, NULL user_defined_type_schema, NULL user_defined_type_name, 'NO' is_insertable_into, 'NO' is_typed, NULL commit_action FROM duckdb_views;"},
     {nullptr, nullptr, nullptr}};
 
 static unique_ptr<CreateViewInfo> GetDefaultView(ClientContext &context, const string &input_schema, const string &input_name) {