duckdb · Mytherin · May 24, 2023 · May 18, 2023 · May 20, 2023 · May 20, 2023
diff --git a/.clang-tidy b/.clang-tidy
@@ -1,4 +1,4 @@
-Checks:          '-*,clang-diagnostic-*,bugprone-*,performance-*,google-explicit-constructor,google-build-using-namespace,google-runtime-int,misc-definitions-in-headers,modernize-use-nullptr,modernize-use-override,-bugprone-macro-parentheses,readability-braces-around-statements,-bugprone-branch-clone,readability-identifier-naming,hicpp-exception-baseclass,misc-throw-by-value-catch-by-reference,-bugprone-signed-char-misuse,-bugprone-misplaced-widening-cast,-bugprone-sizeof-expression,-bugprone-narrowing-conversions,-bugprone-easily-swappable-parameters,google-global-names-in-headers,llvm-header-guard,misc-definitions-in-headers,modernize-use-emplace,modernize-use-bool-literals,-performance-inefficient-string-concatenation,-performance-no-int-to-ptr,readability-container-size-empty'
+Checks:          '-*,clang-diagnostic-*,bugprone-*,performance-*,google-explicit-constructor,google-build-using-namespace,google-runtime-int,misc-definitions-in-headers,modernize-use-nullptr,modernize-use-override,-bugprone-macro-parentheses,readability-braces-around-statements,-bugprone-branch-clone,readability-identifier-naming,hicpp-exception-baseclass,misc-throw-by-value-catch-by-reference,-bugprone-signed-char-misuse,-bugprone-misplaced-widening-cast,-bugprone-sizeof-expression,-bugprone-narrowing-conversions,-bugprone-easily-swappable-parameters,google-global-names-in-headers,llvm-header-guard,misc-definitions-in-headers,modernize-use-emplace,modernize-use-bool-literals,-performance-inefficient-string-concatenation,-performance-no-int-to-ptr,readability-container-size-empty,cppcoreguidelines-pro-type-cstyle-cast'
 WarningsAsErrors: '*'
 HeaderFilterRegex: '.*^(re2.h)'
 AnalyzeTemporaryDtors: false

diff --git a/.github/config/uncovered_files.csv b/.github/config/uncovered_files.csv
@@ -30,7 +30,7 @@ common/exception_format_value.cpp	8
 common/field_writer.cpp	3
 common/file_buffer.cpp	5
 common/file_system.cpp	19
-common/fsst.cpp	5
+common/fsst.cpp	6
 common/gzip_file_system.cpp	40
 common/hive_partitioning.cpp	54
 common/local_file_system.cpp	16
@@ -53,7 +53,7 @@ common/serializer/buffered_deserializer.cpp	3
 common/serializer/buffered_file_reader.cpp	9
 common/serializer/buffered_file_writer.cpp	2
 common/sort/comparators.cpp	72
-common/sort/merge_sorter.cpp	99
+common/sort/merge_sorter.cpp	100
 common/sort/partition_state.cpp	43
 common/sort/sort_state.cpp	10
 common/sort/sorted_block.cpp	20
@@ -258,7 +258,7 @@ function/scalar_function.cpp	7
 function/table/arrow.cpp	61
 function/table/arrow_conversion.cpp	225
 function/table/checkpoint.cpp	7
-function/table/copy_csv.cpp	14
+function/table/copy_csv.cpp	16
 function/table/pragma_detailed_profiling_output.cpp	86
 function/table/pragma_last_profiling_output.cpp	51
 function/table/range.cpp	5
@@ -390,7 +390,7 @@ include/duckdb/main/connection_manager.hpp	3
 include/duckdb/main/database.hpp	2
 include/duckdb/main/prepared_statement.hpp	9
 include/duckdb/main/query_profiler.hpp	9
-include/duckdb/main/query_result.hpp	9
+include/duckdb/main/query_result.hpp	11
 include/duckdb/main/relation.hpp	3
 include/duckdb/main/relation/subquery_relation.hpp	3
 include/duckdb/main/relation/write_parquet_relation.hpp	3
@@ -564,7 +564,7 @@ parser/statement/multi_statement.cpp	7
 parser/statement/select_statement.cpp	2
 parser/statement/set_statement.cpp	3
 parser/statement/update_statement.cpp	2
-parser/tableref.cpp	8
+parser/tableref.cpp	9
 parser/tableref/basetableref.cpp	2
 parser/tableref/emptytableref.cpp	3
 parser/tableref/expressionlistref.cpp	5

diff --git a/extension/autocomplete/sql_auto_complete-extension.cpp b/extension/autocomplete/sql_auto_complete-extension.cpp
@@ -33,7 +33,7 @@ struct SQLAutoCompleteData : public GlobalTableFunctionState {
 };
 
 struct AutoCompleteCandidate {
-	AutoCompleteCandidate(string candidate_p, int32_t score_bonus = 0)
+	explicit AutoCompleteCandidate(string candidate_p, int32_t score_bonus = 0)
 	    : candidate(std::move(candidate_p)), score_bonus(score_bonus) {
 	}
 
@@ -408,7 +408,7 @@ void SQLAutoCompleteExtension::Load(DuckDB &db) {
 }
 
 std::string SQLAutoCompleteExtension::Name() {
-	return "sql_auto_complete";
+	return "autocomplete";
 }
 
 } // namespace duckdb

diff --git a/extension/icu/icu-datepart.cpp b/extension/icu/icu-datepart.cpp
@@ -329,7 +329,7 @@ struct ICUDatePart : public ICUDateFunc {
 			input.ToUnifiedFormat(count, rdata);
 
 			const auto &arg_valid = rdata.validity;
-			auto tdata = (const INPUT_TYPE *)rdata.data;
+			auto tdata = UnifiedVectorFormat::GetData<INPUT_TYPE>(rdata);
 
 			result.SetVectorType(VectorType::FLAT_VECTOR);
 			auto &child_entries = StructVector::GetEntries(result);

diff --git a/extension/inet/inet_functions.cpp b/extension/inet/inet_functions.cpp
@@ -21,7 +21,7 @@ bool INetFunctions::CastVarcharToINET(Vector &source, Vector &result, idx_t coun
 	auto address_data = FlatVector::GetData<hugeint_t>(*entries[1]);
 	auto mask_data = FlatVector::GetData<uint16_t>(*entries[2]);
 
-	auto input = (string_t *)vdata.data;
+	auto input = UnifiedVectorFormat::GetData<string_t>(vdata);
 	bool success = true;
 	for (idx_t i = 0; i < (constant ? 1 : count); i++) {
 		auto idx = vdata.sel->get_index(i);

diff --git a/extension/jemalloc/jemalloc-extension.cpp b/extension/jemalloc/jemalloc-extension.cpp
@@ -15,7 +15,7 @@ std::string JEMallocExtension::Name() {
 }
 
 data_ptr_t JEMallocExtension::Allocate(PrivateAllocatorData *private_data, idx_t size) {
-	return (data_ptr_t)duckdb_jemalloc::je_malloc(size);
+	return data_ptr_cast(duckdb_jemalloc::je_malloc(size));
 }
 
 void JEMallocExtension::Free(PrivateAllocatorData *private_data, data_ptr_t pointer, idx_t size) {
@@ -24,7 +24,7 @@ void JEMallocExtension::Free(PrivateAllocatorData *private_data, data_ptr_t poin
 
 data_ptr_t JEMallocExtension::Reallocate(PrivateAllocatorData *private_data, data_ptr_t pointer, idx_t old_size,
                                          idx_t size) {
-	return (data_ptr_t)duckdb_jemalloc::je_realloc(pointer, size);
+	return data_ptr_cast(duckdb_jemalloc::je_realloc(pointer, size));
 }
 
 } // namespace duckdb

diff --git a/extension/json/include/json_common.hpp b/extension/json/include/json_common.hpp
@@ -38,7 +38,7 @@ class JSONAllocator {
 
 	static inline void *Reallocate(void *ctx, void *ptr, size_t old_size, size_t size) {
 		auto alloc = (ArenaAllocator *)ctx;
-		return alloc->ReallocateAligned((data_ptr_t)ptr, old_size, size);
+		return alloc->ReallocateAligned(data_ptr_cast(ptr), old_size, size);
 	}
 
 	static inline void Free(void *ctx, void *ptr) {

diff --git a/extension/json/include/json_executors.hpp b/extension/json/include/json_executors.hpp
@@ -89,7 +89,7 @@ struct JSONExecutors {
 		UnifiedVectorFormat input_data;
 		auto &input_vector = args.data[0];
 		input_vector.ToUnifiedFormat(count, input_data);
-		auto inputs = (string_t *)input_data.data;
+		auto inputs = UnifiedVectorFormat::GetData<string_t>(input_data);
 
 		ListVector::Reserve(result, list_size);
 		auto list_entries = FlatVector::GetData<list_entry_t>(result);

diff --git a/extension/json/json_functions/json_create.cpp b/extension/json/json_functions/json_create.cpp
@@ -225,7 +225,7 @@ static void AddKeyValuePairs(yyjson_mut_doc *doc, yyjson_mut_val *objs[], Vector
                              idx_t count) {
 	UnifiedVectorFormat key_data;
 	key_v.ToUnifiedFormat(count, key_data);
-	auto keys = (string_t *)key_data.data;
+	auto keys = UnifiedVectorFormat::GetData<string_t>(key_data);
 
 	for (idx_t i = 0; i < count; i++) {
 		auto key_idx = key_data.sel->get_index(i);
@@ -253,7 +253,7 @@ template <class INPUT_TYPE, class TARGET_TYPE>
 static void TemplatedCreateValues(yyjson_mut_doc *doc, yyjson_mut_val *vals[], Vector &value_v, idx_t count) {
 	UnifiedVectorFormat value_data;
 	value_v.ToUnifiedFormat(count, value_data);
-	auto values = (INPUT_TYPE *)value_data.data;
+	auto values = UnifiedVectorFormat::GetData<INPUT_TYPE>(value_data);
 
 	const auto type_is_json = JSONCommon::LogicalTypeIsJSON(value_v.GetType());
 	for (idx_t i = 0; i < count; i++) {
@@ -311,7 +311,7 @@ static void CreateValuesMap(const StructNames &names, yyjson_mut_doc *doc, yyjso
 	// Add the key/value pairs to the values
 	UnifiedVectorFormat map_data;
 	value_v.ToUnifiedFormat(count, map_data);
-	auto map_key_list_entries = (list_entry_t *)map_data.data;
+	auto map_key_list_entries = UnifiedVectorFormat::GetData<list_entry_t>(map_data);
 	for (idx_t i = 0; i < count; i++) {
 		idx_t idx = map_data.sel->get_index(i);
 		if (!map_data.validity.RowIsValid(idx)) {
@@ -358,14 +358,14 @@ static void CreateValuesUnion(const StructNames &names, yyjson_mut_doc *doc, yyj
 		// and the rows where the member is not matching the tag
 		UnifiedVectorFormat key_data;
 		member_key_v.ToUnifiedFormat(count, key_data);
-		auto keys = (string_t *)key_data.data;
+		auto keys = UnifiedVectorFormat::GetData<string_t>(key_data);
 
 		for (idx_t i = 0; i < count; i++) {
 			auto tag_idx = tag_data.sel->get_index(i);
 			if (!tag_data.validity.RowIsValid(tag_idx)) {
 				continue;
 			}
-			auto tag = ((uint8_t *)tag_data.data)[tag_idx];
+			auto tag = (UnifiedVectorFormat::GetData<uint8_t>(tag_data))[tag_idx];
 			if (tag != member_idx) {
 				continue;
 			}
@@ -390,7 +390,7 @@ static void CreateValuesList(const StructNames &names, yyjson_mut_doc *doc, yyjs
 	// Now we add the values to the appropriate JSON arrays
 	UnifiedVectorFormat list_data;
 	value_v.ToUnifiedFormat(count, list_data);
-	auto list_entries = (list_entry_t *)list_data.data;
+	auto list_entries = UnifiedVectorFormat::GetData<list_entry_t>(list_data);
 	for (idx_t i = 0; i < count; i++) {
 		idx_t idx = list_data.sel->get_index(i);
 		if (!list_data.validity.RowIsValid(idx)) {

diff --git a/extension/json/json_functions/json_merge_patch.cpp b/extension/json/json_functions/json_merge_patch.cpp
@@ -36,7 +36,7 @@ static inline void ReadObjects(yyjson_mut_doc *doc, Vector &input, yyjson_mut_va
 	UnifiedVectorFormat input_data;
 	auto &input_vector = input;
 	input_vector.ToUnifiedFormat(count, input_data);
-	auto inputs = (string_t *)input_data.data;
+	auto inputs = UnifiedVectorFormat::GetData<string_t>(input_data);
 
 	// Read the documents
 	for (idx_t i = 0; i < count; i++) {

diff --git a/extension/json/json_functions/json_transform.cpp b/extension/json/json_functions/json_transform.cpp
@@ -776,7 +776,7 @@ static bool TransformFunctionInternal(Vector &input, const idx_t count, Vector &
                                       JSONTransformOptions &options) {
 	UnifiedVectorFormat input_data;
 	input.ToUnifiedFormat(count, input_data);
-	auto inputs = (string_t *)input_data.data;
+	auto inputs = UnifiedVectorFormat::GetData<string_t>(input_data);
 
 	// Read documents
 	auto docs = (yyjson_doc **)alc->malloc(alc->ctx, sizeof(yyjson_doc *) * count);