duckdb 0.6.2-dev2085.0 → 0.6.2-dev2091.0

package/package.json

@@ -2,7 +2,7 @@
   "name": "duckdb",
   "main": "./lib/duckdb.js",
   "types": "./lib/duckdb.d.ts",
-  "version": "0.6.2-dev2085.0",
+  "version": "0.6.2-dev2091.0",
   "description": "DuckDB node.js API",
   "gypfile": true,
   "dependencies": {

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

@@ -211,6 +211,19 @@ void ParquetWriter::SetSchemaProperties(const LogicalType &duckdb_type,
 	}
 }
 
+void VerifyUniqueNames(const vector<string> &names) {
+#ifdef DEBUG
+	unordered_set<string> name_set;
+	name_set.reserve(names.size());
+	for (auto &column : names) {
+		auto res = name_set.insert(column);
+		D_ASSERT(res.second == true);
+	}
+	// If there would be duplicates, these sizes would differ
+	D_ASSERT(name_set.size() == names.size());
+#endif
+}
+
 ParquetWriter::ParquetWriter(FileSystem &fs, string file_name_p, FileOpener *file_opener_p, vector<LogicalType> types_p,
                              vector<string> names_p, CompressionCodec::type codec)
     : file_name(std::move(file_name_p)), sql_types(std::move(types_p)), column_names(std::move(names_p)), codec(codec) {
@@ -237,10 +250,13 @@ ParquetWriter::ParquetWriter(FileSystem &fs, string file_name_p, FileOpener *fil
 	file_meta_data.schema[0].repetition_type = duckdb_parquet::format::FieldRepetitionType::REQUIRED;
 	file_meta_data.schema[0].__isset.repetition_type = true;
 
+	auto &unique_names = column_names;
+	VerifyUniqueNames(unique_names);
+
 	vector<string> schema_path;
 	for (idx_t i = 0; i < sql_types.size(); i++) {
 		column_writers.push_back(ColumnWriter::CreateWriterRecursive(file_meta_data.schema, *this, sql_types[i],
-		                                                             column_names[i], schema_path));
+		                                                             unique_names[i], schema_path));
 	}
 }
 

package/src/duckdb/src/function/table/version/pragma_version.cpp

@@ -1,8 +1,8 @@
 #ifndef DUCKDB_VERSION
-#define DUCKDB_VERSION "0.6.2-dev2085"
+#define DUCKDB_VERSION "0.6.2-dev2091"
 #endif
 #ifndef DUCKDB_SOURCE_ID
-#define DUCKDB_SOURCE_ID "a038924fb3"
+#define DUCKDB_SOURCE_ID "4d9705ca54"
 #endif
 #include "duckdb/function/table/system_functions.hpp"
 #include "duckdb/main/database.hpp"

package/src/duckdb/src/planner/binder/statement/bind_copy.cpp

@@ -33,6 +33,34 @@ static vector<idx_t> ColumnListToIndices(const vector<bool> &vec) {
 	return ret;
 }
 
+vector<string> GetUniqueNames(const vector<string> &original_names) {
+	unordered_set<string> name_set;
+	vector<string> unique_names;
+	unique_names.reserve(original_names.size());
+
+	for (auto &name : original_names) {
+		auto insert_result = name_set.insert(name);
+		if (insert_result.second == false) {
+			// Could not be inserted, name already exists
+			idx_t index = 1;
+			string postfixed_name;
+			while (true) {
+				postfixed_name = StringUtil::Format("%s:%d", name, index);
+				auto res = name_set.insert(postfixed_name);
+				if (!res.second) {
+					index++;
+					continue;
+				}
+				break;
+			}
+			unique_names.push_back(postfixed_name);
+		} else {
+			unique_names.push_back(name);
+		}
+	}
+	return unique_names;
+}
+
 BoundStatement Binder::BindCopyTo(CopyStatement &stmt) {
 	// COPY TO a file
 	auto &config = DBConfig::GetConfig(context);
@@ -99,8 +127,10 @@ BoundStatement Binder::BindCopyTo(CopyStatement &stmt) {
 		use_tmp_file = is_file_and_exists && !per_thread_output && partition_cols.empty() && !is_stdout;
 	}
 
+	auto unique_column_names = GetUniqueNames(select_node.names);
+
 	auto function_data =
-	    copy_function->function.copy_to_bind(context, *stmt.info, select_node.names, select_node.types);
+	    copy_function->function.copy_to_bind(context, *stmt.info, unique_column_names, select_node.types);
 	// now create the copy information
 	auto copy = make_unique<LogicalCopyToFile>(copy_function->function, std::move(function_data));
 	copy->file_path = stmt.info->file_path;
@@ -110,7 +140,8 @@ BoundStatement Binder::BindCopyTo(CopyStatement &stmt) {
 	copy->per_thread_output = per_thread_output;
 	copy->partition_output = !partition_cols.empty();
 	copy->partition_columns = std::move(partition_cols);
-	copy->names = select_node.names;
+
+	copy->names = unique_column_names;
 	copy->expected_types = select_node.types;
 
 	copy->AddChild(std::move(select_node.plan));