duckdb 0.8.1-dev96.0 → 0.8.1

package/binding.gyp

@@ -237,18 +237,18 @@
                 "src/duckdb/third_party/zstd/compress/zstd_lazy.cpp", 
                 "src/duckdb/third_party/zstd/compress/zstd_ldm.cpp", 
                 "src/duckdb/third_party/zstd/compress/zstd_opt.cpp", 
-                "src/duckdb/extension/icu/./icu-dateadd.cpp", 
-                "src/duckdb/extension/icu/./icu-datetrunc.cpp", 
-                "src/duckdb/extension/icu/./icu-datesub.cpp", 
-                "src/duckdb/extension/icu/./icu-table-range.cpp", 
-                "src/duckdb/extension/icu/./icu-timebucket.cpp", 
                 "src/duckdb/extension/icu/./icu-list-range.cpp", 
+                "src/duckdb/extension/icu/./icu-datefunc.cpp", 
                 "src/duckdb/extension/icu/./icu-datepart.cpp", 
-                "src/duckdb/extension/icu/./icu-timezone.cpp", 
+                "src/duckdb/extension/icu/./icu-datetrunc.cpp", 
+                "src/duckdb/extension/icu/./icu-table-range.cpp", 
+                "src/duckdb/extension/icu/./icu-dateadd.cpp", 
                 "src/duckdb/extension/icu/./icu-extension.cpp", 
-                "src/duckdb/extension/icu/./icu-makedate.cpp", 
-                "src/duckdb/extension/icu/./icu-datefunc.cpp", 
                 "src/duckdb/extension/icu/./icu-strptime.cpp", 
+                "src/duckdb/extension/icu/./icu-datesub.cpp", 
+                "src/duckdb/extension/icu/./icu-makedate.cpp", 
+                "src/duckdb/extension/icu/./icu-timezone.cpp", 
+                "src/duckdb/extension/icu/./icu-timebucket.cpp", 
                 "src/duckdb/ub_extension_icu_third_party_icu_common.cpp", 
                 "src/duckdb/ub_extension_icu_third_party_icu_i18n.cpp", 
                 "src/duckdb/extension/icu/third_party/icu/stubdata/stubdata.cpp", 

package/package.json

@@ -2,7 +2,7 @@
   "name": "duckdb",
   "main": "./lib/duckdb.js",
   "types": "./lib/duckdb.d.ts",
-  "version": "0.8.1-dev96.0",
+  "version": "0.8.1",
   "description": "DuckDB node.js API",
   "gypfile": true,
   "dependencies": {
@@ -28,11 +28,13 @@
   },
   "devDependencies": {
     "@types/chai": "^4.3.4",
+    "@types/chai-as-promised": "^7.1.5",
     "@types/mocha": "^10.0.0",
     "@types/node": "^18.11.0",
     "apache-arrow": "^9.0.0",
     "aws-sdk": "^2.790.0",
     "chai": "^4.3.6",
+    "chai-as-promised": "^7.1.1",
     "jsdoc3-parser": "^2.0.0",
     "mocha": "^8.3.0",
     "ts-node": "^10.9.1",

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

@@ -432,7 +432,7 @@ struct ICUDatePart : public ICUDateFunc {
 		throw NotImplementedException("FIXME: serialize icu-datepart");
 	}
 
-	static duckdb::unique_ptr<FunctionData> DeserializeFunction(ClientContext &context, FieldReader &reader,
+	static duckdb::unique_ptr<FunctionData> DeserializeFunction(PlanDeserializationState &state, FieldReader &reader,
 	                                                            ScalarFunction &bound_function) {
 		throw NotImplementedException("FIXME: serialize icu-datepart");
 	}

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

@@ -145,7 +145,7 @@ static void ICUCollateSerialize(FieldWriter &writer, const FunctionData *bind_da
 	throw NotImplementedException("FIXME: serialize icu-collate");
 }
 
-static duckdb::unique_ptr<FunctionData> ICUCollateDeserialize(ClientContext &context, FieldReader &reader,
+static duckdb::unique_ptr<FunctionData> ICUCollateDeserialize(PlanDeserializationState &state, FieldReader &reader,
                                                               ScalarFunction &bound_function) {
 	throw NotImplementedException("FIXME: serialize icu-collate");
 }

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

@@ -1,3 +1,4 @@
+#include "duckdb/common/operator/cast_operators.hpp"
 #include "duckdb/common/types/date.hpp"
 #include "duckdb/common/types/time.hpp"
 #include "duckdb/common/types/timestamp.hpp"
@@ -36,8 +37,8 @@ struct ICUMakeDate : public ICUDateFunc {
 
 	static bool CastToDate(Vector &source, Vector &result, idx_t count, CastParameters &parameters) {
 		auto &cast_data = parameters.cast_data->Cast<CastData>();
-		auto info = (BindData *)cast_data.info.get();
-		CalendarPtr calendar(info->calendar->clone());
+		auto &info = cast_data.info->Cast<BindData>();
+		CalendarPtr calendar(info.calendar->clone());
 
 		UnaryExecutor::Execute<timestamp_t, date_t>(
 		    source, result, count, [&](timestamp_t input) { return Operation(calendar.get(), input); });
@@ -67,7 +68,7 @@ struct ICUMakeTimestampTZFunc : public ICUDateFunc {
 	static inline timestamp_t Operation(icu::Calendar *calendar, T yyyy, T mm, T dd, T hr, T mn, double ss) {
 		const auto year = yyyy + (yyyy < 0);
 
-		const auto secs = int32_t(ss);
+		const auto secs = Cast::Operation<double, int32_t>(ss);
 		ss -= secs;
 		ss *= Interval::MSECS_PER_SEC;
 		const auto millis = int32_t(ss);
@@ -87,7 +88,7 @@ struct ICUMakeTimestampTZFunc : public ICUDateFunc {
 	template <typename T>
 	static void Execute(DataChunk &input, ExpressionState &state, Vector &result) {
 		auto &func_expr = state.expr.Cast<BoundFunctionExpression>();
-		auto &info = (BindData &)*func_expr.bind_info;
+		auto &info = func_expr.bind_info->Cast<BindData>();
 		CalendarPtr calendar_ptr(info.calendar->clone());
 		auto calendar = calendar_ptr.get();
 

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

@@ -53,7 +53,7 @@ struct ICUStrptime : public ICUDateFunc {
 			throw NotImplementedException("FIXME: serialize icu-strptime");
 		}
 
-		static duckdb::unique_ptr<FunctionData> Deserialize(ClientContext &context, FieldReader &reader,
+		static duckdb::unique_ptr<FunctionData> Deserialize(PlanDeserializationState &state, FieldReader &reader,
 		                                                    ScalarFunction &bound_function) {
 			throw NotImplementedException("FIXME: serialize icu-strptime");
 		}

package/src/duckdb/extension/icu/third_party/icu/i18n/nfsubs.cpp

@@ -1310,10 +1310,8 @@ NumeratorSubstitution::doParse(const UnicodeString& text,
         // compute the 'effective' base and prescale the value down
         int64_t n = result.getLong(status); // force conversion!
         int64_t d = 1;
-        int32_t pow = 0;
         while (d <= n) {
             d *= 10;
-            ++pow;
         }
         // now add the zeros
         while (zeroCount > 0) {

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

@@ -30,6 +30,10 @@ JSONFileHandle::JSONFileHandle(unique_ptr<FileHandle> file_handle_p, Allocator &
       requested_reads(0), actual_reads(0), cached_size(0) {
 }
 
+bool JSONFileHandle::IsOpen() const {
+	return file_handle != nullptr;
+}
+
 void JSONFileHandle::Close() {
 	if (file_handle) {
 		file_handle->Close();
@@ -62,19 +66,20 @@ idx_t JSONFileHandle::GetPositionAndSize(idx_t &position, idx_t requested_size)
 	if (actual_size != 0) {
 		requested_reads++;
 	}
+
 	return actual_size;
 }
 
-void JSONFileHandle::ReadAtPosition(const char *pointer, idx_t size, idx_t position, bool sample_run) {
+void JSONFileHandle::ReadAtPosition(char *pointer, idx_t size, idx_t position, bool sample_run) {
 	D_ASSERT(size != 0);
 	if (plain_file_source) {
-		file_handle->Read((void *)pointer, size, position);
+		file_handle->Read(pointer, size, position);
 		actual_reads++;
 		return;
 	}
 
 	if (sample_run) { // Cache the buffer
-		file_handle->Read((void *)pointer, size, position);
+		file_handle->Read(pointer, size, position);
 		actual_reads++;
 		cached_buffers.emplace_back(allocator.Allocate(size));
 		memcpy(cached_buffers.back().get(), pointer, size);
@@ -87,12 +92,12 @@ void JSONFileHandle::ReadAtPosition(const char *pointer, idx_t size, idx_t posit
 		actual_reads++;
 	}
 	if (size != 0) {
-		file_handle->Read((void *)pointer, size, position);
+		file_handle->Read(pointer, size, position);
 		actual_reads++;
 	}
 }
 
-idx_t JSONFileHandle::Read(const char *pointer, idx_t requested_size, bool sample_run) {
+idx_t JSONFileHandle::Read(char *pointer, idx_t requested_size, bool sample_run) {
 	D_ASSERT(requested_size != 0);
 	if (plain_file_source) {
 		auto actual_size = ReadInternal(pointer, requested_size);
@@ -121,7 +126,7 @@ idx_t JSONFileHandle::Read(const char *pointer, idx_t requested_size, bool sampl
 	return actual_size;
 }
 
-idx_t JSONFileHandle::ReadFromCache(const char *&pointer, idx_t &size, idx_t &position) {
+idx_t JSONFileHandle::ReadFromCache(char *&pointer, idx_t &size, idx_t &position) {
 	idx_t read_size = 0;
 	idx_t total_offset = 0;
 
@@ -134,7 +139,7 @@ idx_t JSONFileHandle::ReadFromCache(const char *&pointer, idx_t &size, idx_t &po
 		if (position < total_offset + cached_buffer.GetSize()) {
 			idx_t within_buffer_offset = position - total_offset;
 			idx_t copy_size = MinValue<idx_t>(size, cached_buffer.GetSize() - within_buffer_offset);
-			memcpy((void *)pointer, cached_buffer.get() + within_buffer_offset, copy_size);
+			memcpy(pointer, cached_buffer.get() + within_buffer_offset, copy_size);
 
 			read_size += copy_size;
 			pointer += copy_size;
@@ -147,11 +152,11 @@ idx_t JSONFileHandle::ReadFromCache(const char *&pointer, idx_t &size, idx_t &po
 	return read_size;
 }
 
-idx_t JSONFileHandle::ReadInternal(const char *pointer, const idx_t requested_size) {
+idx_t JSONFileHandle::ReadInternal(char *pointer, const idx_t requested_size) {
 	// Deal with reading from pipes
 	idx_t total_read_size = 0;
 	while (total_read_size < requested_size) {
-		auto read_size = file_handle->Read((void *)(pointer + total_read_size), requested_size - total_read_size);
+		auto read_size = file_handle->Read(pointer + total_read_size, requested_size - total_read_size);
 		if (read_size == 0) {
 			break;
 		}
@@ -165,6 +170,7 @@ BufferedJSONReader::BufferedJSONReader(ClientContext &context, BufferedJSONReade
 }
 
 void BufferedJSONReader::OpenJSONFile() {
+	D_ASSERT(!IsDone());
 	lock_guard<mutex> guard(lock);
 	auto &file_system = FileSystem::GetFileSystem(context);
 	auto regular_file_handle =
@@ -186,6 +192,13 @@ bool BufferedJSONReader::IsOpen() const {
 	return file_handle != nullptr;
 }
 
+bool BufferedJSONReader::IsDone() const {
+	if (file_handle) {
+		return !file_handle->IsOpen();
+	}
+	return false;
+}
+
 BufferedJSONReaderOptions &BufferedJSONReader::GetOptions() {
 	return options;
 }
@@ -212,10 +225,6 @@ void BufferedJSONReader::SetRecordType(duckdb::JSONRecordType type) {
 	options.record_type = type;
 }
 
-bool BufferedJSONReader::IsParallel() const {
-	return options.format == JSONFormat::NEWLINE_DELIMITED && file_handle->CanSeek();
-}
-
 const string &BufferedJSONReader::GetFileName() const {
 	return file_name;
 }
@@ -288,7 +297,7 @@ void BufferedJSONReader::ThrowTransformError(idx_t buf_index, idx_t line_or_obje
                                              const string &error_message) {
 	string unit = options.format == JSONFormat::NEWLINE_DELIMITED ? "line" : "record/value";
 	auto line = GetLineNumber(buf_index, line_or_object_in_buf);
-	throw InvalidInputException("JSON transform error in file \"%s\", in %s %llu: %s.", file_name, unit, line,
+	throw InvalidInputException("JSON transform error in file \"%s\", in %s %llu: %s", file_name, unit, line,
 	                            error_message);
 }
 

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

@@ -71,6 +71,7 @@ public:
 struct JSONFileHandle {
 public:
 	JSONFileHandle(unique_ptr<FileHandle> file_handle, Allocator &allocator);
+	bool IsOpen() const;
 	void Close();
 
 	idx_t FileSize() const;
@@ -80,15 +81,15 @@ public:
 	void Seek(idx_t position);
 
 	idx_t GetPositionAndSize(idx_t &position, idx_t requested_size);
-	void ReadAtPosition(const char *pointer, idx_t size, idx_t position, bool sample_run);
-	idx_t Read(const char *pointer, idx_t requested_size, bool sample_run);
+	void ReadAtPosition(char *pointer, idx_t size, idx_t position, bool sample_run);
+	idx_t Read(char *pointer, idx_t requested_size, bool sample_run);
 
 	void Reset();
 	bool RequestedReadsComplete();
 
 private:
-	idx_t ReadFromCache(const char *&pointer, idx_t &size, idx_t &position);
-	idx_t ReadInternal(const char *pointer, const idx_t requested_size);
+	idx_t ReadFromCache(char *&pointer, idx_t &size, idx_t &position);
+	idx_t ReadInternal(char *pointer, const idx_t requested_size);
 
 private:
 	//! The JSON file handle
@@ -139,6 +140,7 @@ public:
 	void OpenJSONFile();
 	void CloseJSONFile();
 	bool IsOpen() const;
+	bool IsDone() const;
 
 	BufferedJSONReaderOptions &GetOptions();
 	const BufferedJSONReaderOptions &GetOptions() const;
@@ -148,8 +150,6 @@ public:
 	JSONRecordType GetRecordType() const;
 	void SetRecordType(JSONRecordType type);
 
-	bool IsParallel() const;
-
 	const string &GetFileName() const;
 	JSONFileHandle &GetFileHandle() const;
 

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

@@ -170,6 +170,16 @@ public:
 	}
 
 public:
+	template <class T>
+	static T *AllocateArray(yyjson_alc *alc, idx_t count) {
+		return reinterpret_cast<T *>(alc->malloc(alc->ctx, sizeof(T) * count));
+	}
+
+	template <class T>
+	static T *AllocateArray(yyjson_mut_doc *doc, idx_t count) {
+		return AllocateArray<T>(&doc->alc, count);
+	}
+
 	static inline yyjson_mut_doc *CreateDocument(yyjson_alc *alc) {
 		D_ASSERT(alc);
 		return yyjson_mut_doc_new(alc);
@@ -419,11 +429,11 @@ private:
 
 template <>
 inline char *JSONCommon::WriteVal(yyjson_val *val, yyjson_alc *alc, idx_t &len) {
-	return yyjson_val_write_opts(val, JSONCommon::WRITE_FLAG, alc, (size_t *)&len, nullptr);
+	return yyjson_val_write_opts(val, JSONCommon::WRITE_FLAG, alc, reinterpret_cast<size_t *>(&len), nullptr);
 }
 template <>
 inline char *JSONCommon::WriteVal(yyjson_mut_val *val, yyjson_alc *alc, idx_t &len) {
-	return yyjson_mut_val_write_opts(val, JSONCommon::WRITE_FLAG, alc, (size_t *)&len, nullptr);
+	return yyjson_mut_val_write_opts(val, JSONCommon::WRITE_FLAG, alc, reinterpret_cast<size_t *>(&len), nullptr);
 }
 
 template <>

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

@@ -232,6 +232,8 @@ private:
 	void ThrowObjectSizeError(const idx_t object_size);
 	void ThrowInvalidAtEndError();
 
+	bool IsParallel(JSONScanGlobalState &gstate) const;
+
 private:
 	//! Bind data
 	const JSONScanData &bind_data;
@@ -245,7 +247,7 @@ private:
 	bool is_last;
 
 	//! Current buffer read info
-	const char *buffer_ptr;
+	char *buffer_ptr;
 	idx_t buffer_size;
 	idx_t buffer_offset;
 	idx_t prev_buffer_remainder;
@@ -290,7 +292,8 @@ public:
 	                                  vector<unique_ptr<Expression>> &filters);
 
 	static void Serialize(FieldWriter &writer, const FunctionData *bind_data_p, const TableFunction &function);
-	static unique_ptr<FunctionData> Deserialize(ClientContext &context, FieldReader &reader, TableFunction &function);
+	static unique_ptr<FunctionData> Deserialize(PlanDeserializationState &state, FieldReader &reader,
+	                                            TableFunction &function);
 
 	static void TableFunctionDefaults(TableFunction &table_function);
 };

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

@@ -114,6 +114,11 @@ static void JSONContainsFunction(DataChunk &args, ExpressionState &state, Vector
 	auto &needles = args.data[1];
 
 	if (needles.GetVectorType() == VectorType::CONSTANT_VECTOR) {
+		if (ConstantVector::IsNull(needles)) {
+			result.SetVectorType(VectorType::CONSTANT_VECTOR);
+			ConstantVector::SetNull(result, true);
+			return;
+		}
 		auto &needle_str = *ConstantVector::GetData<string_t>(needles);
 		auto needle_doc = JSONCommon::ReadDocument(needle_str, JSONCommon::READ_FLAG, lstate.json_allocator.GetYYAlc());
 		UnaryExecutor::Execute<string_t, bool>(haystacks, result, args.size(), [&](string_t haystack_str) {

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

@@ -276,7 +276,7 @@ static void CreateValuesStruct(const StructNames &names, yyjson_mut_doc *doc, yy
 		vals[i] = yyjson_mut_obj(doc);
 	}
 	// Initialize re-usable array for the nested values
-	auto nested_vals = (yyjson_mut_val **)doc->alc.malloc(doc->alc.ctx, sizeof(yyjson_mut_val *) * count);
+	auto nested_vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 
 	// Add the key/value pairs to the values
 	auto &entries = StructVector::GetEntries(value_v);
@@ -301,12 +301,12 @@ static void CreateValuesMap(const StructNames &names, yyjson_mut_doc *doc, yyjso
 	// Create nested keys
 	auto &map_key_v = MapVector::GetKeys(value_v);
 	auto map_key_count = ListVector::GetListSize(value_v);
-	auto nested_keys = (yyjson_mut_val **)doc->alc.malloc(doc->alc.ctx, sizeof(yyjson_mut_val *) * map_key_count);
+	auto nested_keys = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, map_key_count);
 	TemplatedCreateValues<string_t, string_t>(doc, nested_keys, map_key_v, map_key_count);
 	// Create nested values
 	auto &map_val_v = MapVector::GetValues(value_v);
 	auto map_val_count = ListVector::GetListSize(value_v);
-	auto nested_vals = (yyjson_mut_val **)doc->alc.malloc(doc->alc.ctx, sizeof(yyjson_mut_val *) * map_val_count);
+	auto nested_vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, map_val_count);
 	CreateValues(names, doc, nested_vals, map_val_v, map_val_count);
 	// Add the key/value pairs to the values
 	UnifiedVectorFormat map_data;
@@ -338,7 +338,7 @@ static void CreateValuesUnion(const StructNames &names, yyjson_mut_doc *doc, yyj
 	}
 
 	// Initialize re-usable array for the nested values
-	auto nested_vals = (yyjson_mut_val **)doc->alc.malloc(doc->alc.ctx, sizeof(yyjson_mut_val *) * count);
+	auto nested_vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 
 	auto &tag_v = UnionVector::GetTags(value_v);
 	UnifiedVectorFormat tag_data;
@@ -384,7 +384,7 @@ static void CreateValuesList(const StructNames &names, yyjson_mut_doc *doc, yyjs
 	// Initialize array for the nested values
 	auto &child_v = ListVector::GetEntry(value_v);
 	auto child_count = ListVector::GetListSize(value_v);
-	auto nested_vals = (yyjson_mut_val **)doc->alc.malloc(doc->alc.ctx, sizeof(yyjson_mut_val *) * child_count);
+	auto nested_vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, child_count);
 	// Fill nested_vals with list values
 	CreateValues(names, doc, nested_vals, child_v, child_count);
 	// Now we add the values to the appropriate JSON arrays
@@ -501,12 +501,12 @@ static void ObjectFunction(DataChunk &args, ExpressionState &state, Vector &resu
 	// Initialize values
 	const idx_t count = args.size();
 	auto doc = JSONCommon::CreateDocument(alc);
-	auto objs = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto objs = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 	for (idx_t i = 0; i < count; i++) {
 		objs[i] = yyjson_mut_obj(doc);
 	}
 	// Initialize a re-usable value array
-	auto vals = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 	// Loop through key/value pairs
 	for (idx_t pair_idx = 0; pair_idx < args.data.size() / 2; pair_idx++) {
 		Vector &key_v = args.data[pair_idx * 2];
@@ -533,12 +533,12 @@ static void ArrayFunction(DataChunk &args, ExpressionState &state, Vector &resul
 	// Initialize arrays
 	const idx_t count = args.size();
 	auto doc = JSONCommon::CreateDocument(alc);
-	auto arrs = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto arrs = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 	for (idx_t i = 0; i < count; i++) {
 		arrs[i] = yyjson_mut_arr(doc);
 	}
 	// Initialize a re-usable value array
-	auto vals = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 	// Loop through args
 	for (auto &v : args.data) {
 		CreateValues(info.const_struct_names, doc, vals, v, count);
@@ -561,7 +561,7 @@ static void ToJSONFunctionInternal(const StructNames &names, Vector &input, cons
                                    yyjson_alc *alc) {
 	// Initialize array for values
 	auto doc = JSONCommon::CreateDocument(alc);
-	auto vals = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto vals = JSONCommon::AllocateArray<yyjson_mut_val *>(doc, count);
 	CreateValues(names, doc, vals, input, count);
 
 	// Write JSON values to string

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

@@ -59,11 +59,11 @@ static void MergePatchFunction(DataChunk &args, ExpressionState &state, Vector &
 	const auto count = args.size();
 
 	// Read the first json arg
-	auto origs = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto origs = JSONCommon::AllocateArray<yyjson_mut_val *>(alc, count);
 	ReadObjects(doc, args.data[0], origs, count);
 
 	// Read the next json args one by one and merge them into the first json arg
-	auto patches = (yyjson_mut_val **)alc->malloc(alc->ctx, sizeof(yyjson_mut_val *) * count);
+	auto patches = JSONCommon::AllocateArray<yyjson_mut_val *>(alc, count);
 	for (idx_t arg_idx = 1; arg_idx < args.data.size(); arg_idx++) {
 		ReadObjects(doc, args.data[arg_idx], patches, count);
 		for (idx_t i = 0; i < count; i++) {

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

@@ -108,7 +108,7 @@ static void JsonSerializeFunction(DataChunk &args, ExpressionState &state, Vecto
 			idx_t len;
 			auto data = yyjson_mut_val_write_opts(result_obj,
 			                                      info.format ? JSONCommon::WRITE_PRETTY_FLAG : JSONCommon::WRITE_FLAG,
-			                                      alc, (size_t *)&len, nullptr);
+			                                      alc, reinterpret_cast<size_t *>(&len), nullptr);
 			if (data == nullptr) {
 				throw SerializationException(
 				    "Failed to serialize json, perhaps the query contains invalid utf8 characters?");
@@ -124,7 +124,7 @@ static void JsonSerializeFunction(DataChunk &args, ExpressionState &state, Vecto
 			idx_t len;
 			auto data = yyjson_mut_val_write_opts(result_obj,
 			                                      info.format ? JSONCommon::WRITE_PRETTY_FLAG : JSONCommon::WRITE_FLAG,
-			                                      alc, (size_t *)&len, nullptr);
+			                                      alc, reinterpret_cast<size_t *>(&len), nullptr);
 			return StringVector::AddString(result, data, len);
 		}
 	});

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

@@ -150,7 +150,8 @@ void JSONStructureNode::RefineCandidateTypesArray(yyjson_val *vals[], idx_t coun
 	}
 
 	idx_t offset = 0;
-	auto child_vals = (yyjson_val **)allocator.AllocateAligned(total_list_size * sizeof(yyjson_val *));
+	auto child_vals =
+	    reinterpret_cast<yyjson_val **>(allocator.AllocateAligned(total_list_size * sizeof(yyjson_val *)));
 
 	size_t idx, max;
 	yyjson_val *child_val;
@@ -173,11 +174,12 @@ void JSONStructureNode::RefineCandidateTypesObject(yyjson_val *vals[], idx_t cou
 	vector<yyjson_val **> child_vals;
 	child_vals.reserve(child_count);
 	for (idx_t child_idx = 0; child_idx < child_count; child_idx++) {
-		child_vals.emplace_back((yyjson_val **)allocator.AllocateAligned(count * sizeof(yyjson_val *)));
+		child_vals.emplace_back(
+		    reinterpret_cast<yyjson_val **>(allocator.AllocateAligned(count * sizeof(yyjson_val *))));
 	}
 
 	idx_t found_key_count;
-	auto found_keys = (bool *)allocator.AllocateAligned(sizeof(bool) * child_count);
+	auto found_keys = reinterpret_cast<bool *>(allocator.AllocateAligned(sizeof(bool) * child_count));
 
 	const auto &key_map = desc.key_map;
 	size_t idx, max;

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

@@ -216,7 +216,7 @@ static inline bool GetValueString(yyjson_val *val, yyjson_alc *alc, string_t &re
 
 template <class T>
 static bool TransformNumerical(yyjson_val *vals[], Vector &result, const idx_t count, JSONTransformOptions &options) {
-	auto data = (T *)FlatVector::GetData(result);
+	auto data = FlatVector::GetData<T>(result);
 	auto &validity = FlatVector::Validity(result);
 
 	bool success = true;
@@ -238,7 +238,7 @@ static bool TransformNumerical(yyjson_val *vals[], Vector &result, const idx_t c
 template <class T>
 static bool TransformDecimal(yyjson_val *vals[], Vector &result, const idx_t count, uint8_t width, uint8_t scale,
                              JSONTransformOptions &options) {
-	auto data = (T *)FlatVector::GetData(result);
+	auto data = FlatVector::GetData<T>(result);
 	auto &validity = FlatVector::Validity(result);
 
 	bool success = true;
@@ -373,7 +373,7 @@ static bool TransformFromStringWithFormat(yyjson_val *vals[], Vector &result, co
 }
 
 static bool TransformToString(yyjson_val *vals[], yyjson_alc *alc, Vector &result, const idx_t count) {
-	auto data = (string_t *)FlatVector::GetData(result);
+	auto data = FlatVector::GetData<string_t>(result);
 	auto &validity = FlatVector::Validity(result);
 	for (idx_t i = 0; i < count; i++) {
 		const auto &val = vals[i];
@@ -400,11 +400,11 @@ bool JSONTransform::TransformObject(yyjson_val *objects[], yyjson_alc *alc, cons
 	nested_vals.reserve(column_count);
 	for (idx_t col_idx = 0; col_idx < column_count; col_idx++) {
 		key_map.insert({{names[col_idx].c_str(), names[col_idx].length()}, col_idx});
-		nested_vals.push_back((yyjson_val **)alc->malloc(alc->ctx, sizeof(yyjson_val *) * count));
+		nested_vals.push_back(JSONCommon::AllocateArray<yyjson_val *>(alc, count));
 	}
 
 	idx_t found_key_count;
-	auto found_keys = (bool *)alc->malloc(alc->ctx, sizeof(bool) * column_count);
+	auto found_keys = JSONCommon::AllocateArray<bool>(alc, column_count);
 
 	bool success = true;
 
@@ -558,7 +558,7 @@ static bool TransformArray(yyjson_val *arrays[], yyjson_alc *alc, Vector &result
 	ListVector::Reserve(result, offset);
 
 	// Initialize array for the nested values
-	auto nested_vals = (yyjson_val **)alc->malloc(alc->ctx, sizeof(yyjson_val *) * offset);
+	auto nested_vals = JSONCommon::AllocateArray<yyjson_val *>(alc, offset);
 
 	// Get array values
 	size_t idx, max;
@@ -617,8 +617,8 @@ static bool TransformObjectToMap(yyjson_val *objects[], yyjson_alc *alc, Vector
 	auto list_entries = FlatVector::GetData<list_entry_t>(result);
 	auto &list_validity = FlatVector::Validity(result);
 
-	auto keys = (yyjson_val **)alc->malloc(alc->ctx, sizeof(yyjson_val *) * list_size);
-	auto vals = (yyjson_val **)alc->malloc(alc->ctx, sizeof(yyjson_val *) * list_size);
+	auto keys = JSONCommon::AllocateArray<yyjson_val *>(alc, list_size);
+	auto vals = JSONCommon::AllocateArray<yyjson_val *>(alc, list_size);
 
 	bool success = true;
 	idx_t list_offset = 0;
@@ -675,7 +675,7 @@ static bool TransformObjectToMap(yyjson_val *objects[], yyjson_alc *alc, Vector
 }
 
 bool TransformToJSON(yyjson_val *vals[], yyjson_alc *alc, Vector &result, const idx_t count) {
-	auto data = (string_t *)FlatVector::GetData(result);
+	auto data = FlatVector::GetData<string_t>(result);
 	auto &validity = FlatVector::Validity(result);
 	for (idx_t i = 0; i < count; i++) {
 		const auto &val = vals[i];
@@ -779,8 +779,8 @@ static bool TransformFunctionInternal(Vector &input, const idx_t count, Vector &
 	auto inputs = UnifiedVectorFormat::GetData<string_t>(input_data);
 
 	// Read documents
-	auto docs = (yyjson_doc **)alc->malloc(alc->ctx, sizeof(yyjson_doc *) * count);
-	auto vals = (yyjson_val **)alc->malloc(alc->ctx, sizeof(yyjson_val *) * count);
+	auto docs = JSONCommon::AllocateArray<yyjson_doc *>(alc, count);
+	auto vals = JSONCommon::AllocateArray<yyjson_val *>(alc, count);
 	auto &result_validity = FlatVector::Validity(result);
 	for (idx_t i = 0; i < count; i++) {
 		auto idx = input_data.sel->get_index(i);

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

@@ -258,7 +258,8 @@ static void ReadJSONFunction(ClientContext &context, TableFunctionInput &data_p,
 			string hint =
 			    gstate.bind_data.auto_detect
 			        ? "\nTry increasing 'sample_size', reducing 'maximum_depth', specifying 'columns', 'format' or "
-			          "'records' manually, or setting 'ignore_errors' to true."
+			          "'records' manually, setting 'ignore_errors' to true, or setting 'union_by_name' to true when "
+			          "reading multiple files with a different structure."
 			        : "\nTry setting 'auto_detect' to true, specifying 'format' or 'records' manually, or setting "
 			          "'ignore_errors' to true.";
 			lstate.ThrowTransformError(lstate.transform_options.object_index,

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

@@ -15,6 +15,9 @@ namespace duckdb {
 static void CheckPath(const Value &path_val, string &path, size_t &len) {
 	string error;
 	Value path_str_val;
+	if (path_val.IsNull()) {
+		throw InvalidInputException("JSON path cannot be NULL");
+	}
 	if (!path_val.DefaultTryCastAs(LogicalType::VARCHAR, path_str_val, &error)) {
 		throw InvalidInputException(error);
 	}
@@ -220,8 +223,8 @@ static bool CastVarcharToJSON(Vector &source, Vector &result, idx_t count, CastP
 	bool success = true;
 	UnaryExecutor::ExecuteWithNulls<string_t, string_t>(
 	    source, result, count, [&](string_t input, ValidityMask &mask, idx_t idx) {
-		    auto data = (char *)(input.GetData());
-		    auto length = input.GetSize();
+		    auto data = input.GetDataWriteable();
+		    const auto length = input.GetSize();
 
 		    yyjson_read_err error;
 		    auto doc = JSONCommon::ReadDocumentUnsafe(data, length, JSONCommon::READ_FLAG, alc, &error);
@@ -236,7 +239,7 @@ static bool CastVarcharToJSON(Vector &source, Vector &result, idx_t count, CastP
 		    }
 		    return input;
 	    });
-	result.Reinterpret(source);
+	StringVector::AddHeapReference(result, source);
 	return success;
 }