duckdb 0.3.2-dev1262.0 → 0.3.2-dev1292.0

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "duckdb",
   "main": "./lib/duckdb.js",
-  "version": "0.3.2-dev1262.0",
+  "version": "0.3.2-dev1292.0",
   "description": "DuckDB node.js API",
   "gypfile": true,
   "dependencies": {

package/src/duckdb.cpp

@@ -58342,8 +58342,8 @@ bool BufferedCSVReader::ReadBuffer(idx_t &start) {
 	while (remaining > buffer_read_size) {
 		buffer_read_size *= 2;
 	}
-	if (remaining + buffer_read_size > MAXIMUM_CSV_LINE_SIZE) {
-		throw InvalidInputException("Maximum line size of %llu bytes exceeded!", MAXIMUM_CSV_LINE_SIZE);
+	if (remaining + buffer_read_size > options.maximum_line_size) {
+		throw InvalidInputException("Maximum line size of %llu bytes exceeded!", options.maximum_line_size);
 	}
 	buffer = unique_ptr<char[]>(new char[buffer_read_size + remaining + 1]);
 	buffer_size = remaining + buffer_read_size;
@@ -91497,6 +91497,8 @@ static bool ParseBaseOption(BufferedCSVReaderOptions &options, string &loption,
 		options.compression = FileCompressionTypeFromString(ParseString(set));
 	} else if (loption == "skip") {
 		options.skip_rows = ParseInteger(set);
+	} else if (loption == "max_line_size" || loption == "maximum_line_size") {
+		options.maximum_line_size = ParseInteger(set);
 	} else {
 		// unrecognized option in base CSV
 		return false;
@@ -92801,6 +92803,8 @@ static unique_ptr<FunctionData> ReadCSVBind(ClientContext &context, vector<Value
 			result->include_file_name = BooleanValue::Get(kv.second);
 		} else if (loption == "skip") {
 			options.skip_rows = kv.second.GetValue<int64_t>();
+		} else if (loption == "max_line_size" || loption == "maximum_line_size") {
+			options.maximum_line_size = kv.second.GetValue<int64_t>();
 		} else {
 			throw InternalException("Unrecognized parameter %s", kv.first);
 		}
@@ -92907,6 +92911,8 @@ static void ReadCSVAddNamedParameters(TableFunction &table_function) {
 	table_function.named_parameters["compression"] = LogicalType::VARCHAR;
 	table_function.named_parameters["filename"] = LogicalType::BOOLEAN;
 	table_function.named_parameters["skip"] = LogicalType::BIGINT;
+	table_function.named_parameters["max_line_size"] = LogicalType::VARCHAR;
+	table_function.named_parameters["maximum_line_size"] = LogicalType::VARCHAR;
 }
 
 double CSVReaderProgress(ClientContext &context, const FunctionData *bind_data_p) {
@@ -167149,8 +167155,8 @@ void UpdateSegment::CleanupUpdate(UpdateInfo *info) {
 //===--------------------------------------------------------------------===//
 // Check for conflicts in update
 //===--------------------------------------------------------------------===//
-static void CheckForConflicts(UpdateInfo *info, Transaction &transaction, row_t *ids, idx_t count, row_t offset,
-                              UpdateInfo *&node) {
+static void CheckForConflicts(UpdateInfo *info, Transaction &transaction, row_t *ids, const SelectionVector &sel,
+                              idx_t count, row_t offset, UpdateInfo *&node) {
 	if (!info) {
 		return;
 	}
@@ -167162,7 +167168,7 @@ static void CheckForConflicts(UpdateInfo *info, Transaction &transaction, row_t
 		// as both ids and info->tuples are sorted, this is similar to a merge join
 		idx_t i = 0, j = 0;
 		while (true) {
-			auto id = ids[i] - offset;
+			auto id = ids[sel.get_index(i)] - offset;
 			if (id == info->tuples[j]) {
 				throw TransactionException("Conflict on update!");
 			} else if (id < info->tuples[j]) {
@@ -167180,7 +167186,7 @@ static void CheckForConflicts(UpdateInfo *info, Transaction &transaction, row_t
 			}
 		}
 	}
-	CheckForConflicts(info->next, transaction, ids, count, offset, node);
+	CheckForConflicts(info->next, transaction, ids, sel, count, offset, node);
 }
 
 //===--------------------------------------------------------------------===//
@@ -167713,7 +167719,7 @@ void UpdateSegment::Update(Transaction &transaction, idx_t column_index, Vector
 		// there is already a version here, check if there are any conflicts and search for the node that belongs to
 		// this transaction in the version chain
 		auto base_info = root->info[vector_index]->info.get();
-		CheckForConflicts(base_info->next, transaction, ids, count, vector_offset, node);
+		CheckForConflicts(base_info->next, transaction, ids, sel, count, vector_offset, node);
 
 		// there are no conflicts
 		// first, check if this thread has already done any updates

package/src/duckdb.hpp

@@ -11,8 +11,8 @@ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLI
 #pragma once
 #define DUCKDB_AMALGAMATION 1
 #define DUCKDB_AMALGAMATION_EXTENDED 1
-#define DUCKDB_SOURCE_ID "477c52aed"
-#define DUCKDB_VERSION "v0.3.2-dev1262"
+#define DUCKDB_SOURCE_ID "e6f479f2f"
+#define DUCKDB_VERSION "v0.3.2-dev1292"
 //===----------------------------------------------------------------------===//
 //                         DuckDB
 //
@@ -20654,6 +20654,9 @@ struct BufferedCSVReaderOptions {
 	idx_t buffer_size = STANDARD_VECTOR_SIZE * 100;
 	//! Consider all columns to be of type varchar
 	bool all_varchar = false;
+	//! Maximum CSV line size: specified because if we reach this amount, we likely have wrong delimiters (default: 2MB)
+	idx_t maximum_line_size = 2097152;
+
 	//! The date format to use (if any is specified)
 	std::map<LogicalTypeId, StrpTimeFormat> date_format = {{LogicalTypeId::DATE, {}}, {LogicalTypeId::TIMESTAMP, {}}};
 	//! Whether or not a type format is specified
@@ -20670,8 +20673,6 @@ enum class ParserMode : uint8_t { PARSING = 0, SNIFFING_DIALECT = 1, SNIFFING_DA
 class BufferedCSVReader {
 	//! Initial buffer read size; can be extended for long lines
 	static constexpr idx_t INITIAL_BUFFER_SIZE = 16384;
-	//! Maximum CSV line size: specified because if we reach this amount, we likely have the wrong delimiters
-	static constexpr idx_t MAXIMUM_CSV_LINE_SIZE = 1048576;
 	ParserMode mode;
 
 public: