duckdb 0.5.1-dev7.0 → 0.5.1-dev72.0

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "duckdb",
   "main": "./lib/duckdb.js",
-  "version": "0.5.1-dev7.0",
+  "version": "0.5.1-dev72.0",
   "description": "DuckDB node.js API",
   "gypfile": true,
   "dependencies": {

package/src/duckdb.cpp

@@ -620,7 +620,88 @@ public:
 
 } // namespace duckdb
 
+//===----------------------------------------------------------------------===//
+//                         DuckDB
+//
+// extension_functions.hpp
+//
+//
+//===----------------------------------------------------------------------===//
+
+
+
+
 
+namespace duckdb {
+
+struct ExtensionFunction {
+	char function[48];
+	char extension[48];
+};
+
+static constexpr ExtensionFunction EXTENSION_FUNCTIONS[] = {
+    {"->>", "json"},
+    {"array_to_json", "json"},
+    {"create_fts_index", "fts"},
+    {"dbgen", "tpch"},
+    {"drop_fts_index", "fts"},
+    {"dsdgen", "tpcds"},
+    {"excel_text", "excel"},
+    {"from_json", "json"},
+    {"from_json_strict", "json"},
+    {"from_substrait", "substrait"},
+    {"get_substrait", "substrait"},
+    {"get_substrait_json", "substrait"},
+    {"icu_calendar_names", "icu"},
+    {"icu_sort_key", "icu"},
+    {"json", "json"},
+    {"json_array", "json"},
+    {"json_array_length", "json"},
+    {"json_extract", "json"},
+    {"json_extract_path", "json"},
+    {"json_extract_path_text", "json"},
+    {"json_extract_string", "json"},
+    {"json_group_array", "json"},
+    {"json_group_object", "json"},
+    {"json_group_structure", "json"},
+    {"json_merge_patch", "json"},
+    {"json_object", "json"},
+    {"json_quote", "json"},
+    {"json_structure", "json"},
+    {"json_transform", "json"},
+    {"json_transform_strict", "json"},
+    {"json_type", "json"},
+    {"json_valid", "json"},
+    {"make_timestamptz", "icu"},
+    {"parquet_metadata", "parquet"},
+    {"parquet_scan", "parquet"},
+    {"parquet_schema", "parquet"},
+    {"pg_timezone_names", "icu"},
+    {"postgres_attach", "postgres_scanner"},
+    {"postgres_scan", "postgres_scanner"},
+    {"postgres_scan_pushdown", "postgres_scanner"},
+    {"read_json_objects", "json"},
+    {"read_ndjson_objects", "json"},
+    {"read_parquet", "parquet"},
+    {"row_to_json", "json"},
+    {"sqlite_attach", "sqlite_scanner"},
+    {"sqlite_scan", "sqlite_scanner"},
+    {"stem", "fts"},
+    {"text", "excel"},
+    {"to_json", "json"},
+    {"tpcds", "tpcds"},
+    {"tpcds_answers", "tpcds"},
+    {"tpcds_queries", "tpcds"},
+    {"tpch", "tpch"},
+    {"tpch_answers", "tpch"},
+    {"tpch_queries", "tpch"},
+    {"visualize_diff_profiling_output", "visualizer"},
+    {"visualize_json_profiling_output", "visualizer"},
+    {"visualize_last_profiling_output", "visualizer"},
+};
+} // namespace duckdb
+
+#include <algorithm>
 namespace duckdb {
 
 string SimilarCatalogEntry::GetQualifiedName() const {
@@ -823,6 +904,16 @@ SimilarCatalogEntry Catalog::SimilarEntryInSchemas(ClientContext &context, const
 	return {most_similar.first, most_similar.second, schema_of_most_similar};
 }
 
+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) {
+		return it->extension;
+	}
+	return "";
+}
 CatalogException Catalog::CreateMissingEntryException(ClientContext &context, const string &entry_name,
                                                       CatalogType type, const vector<SchemaCatalogEntry *> &schemas,
                                                       QueryErrorContext error_context) {
@@ -836,7 +927,12 @@ CatalogException Catalog::CreateMissingEntryException(ClientContext &context, co
 		}
 	});
 	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);
+	}
 	string did_you_mean;
 	if (unseen_entry.Found() && unseen_entry.distance < entry.distance) {
 		did_you_mean = "\nDid you mean \"" + unseen_entry.GetQualifiedName() + "\"?";
@@ -33336,6 +33432,9 @@ void RowOperations::UnswizzleHeapPointer(const RowLayout &layout, const data_ptr
 
 static inline void VerifyUnswizzledString(const RowLayout &layout, const idx_t &col_idx, const data_ptr_t &row_ptr) {
 #ifdef DEBUG
+	if (layout.GetTypes()[col_idx] == LogicalTypeId::BLOB) {
+		return;
+	}
 	idx_t entry_idx;
 	idx_t idx_in_entry;
 	ValidityBytes::GetEntryIndex(col_idx, entry_idx, idx_in_entry);
@@ -35673,7 +35772,10 @@ struct SortConstants {
 
 struct SortLayout {
 public:
+	SortLayout() {
+	}
 	explicit SortLayout(const vector<BoundOrderByNode> &orders);
+	SortLayout GetPrefixComparisonLayout(idx_t num_prefix_cols) const;
 
 public:
 	idx_t column_count;
@@ -37324,6 +37426,32 @@ SortLayout::SortLayout(const vector<BoundOrderByNode> &orders)
 	blob_layout.Initialize(blob_layout_types);
 }
 
+SortLayout SortLayout::GetPrefixComparisonLayout(idx_t num_prefix_cols) const {
+	SortLayout result;
+	result.column_count = num_prefix_cols;
+	result.all_constant = true;
+	result.comparison_size = 0;
+	for (idx_t col_idx = 0; col_idx < num_prefix_cols; col_idx++) {
+		result.order_types.push_back(order_types[col_idx]);
+		result.order_by_null_types.push_back(order_by_null_types[col_idx]);
+		result.logical_types.push_back(logical_types[col_idx]);
+
+		result.all_constant = result.all_constant && constant_size[col_idx];
+		result.constant_size.push_back(constant_size[col_idx]);
+
+		result.comparison_size += column_sizes[col_idx];
+		result.column_sizes.push_back(column_sizes[col_idx]);
+
+		result.prefix_lengths.push_back(prefix_lengths[col_idx]);
+		result.stats.push_back(stats[col_idx]);
+		result.has_null.push_back(has_null[col_idx]);
+	}
+	result.entry_size = entry_size;
+	result.blob_layout = blob_layout;
+	result.sorting_to_blob_col = sorting_to_blob_col;
+	return result;
+}
+
 LocalSortState::LocalSortState() : initialized(false) {
 }
 
@@ -64927,12 +65055,14 @@ public:
 
 	WindowGlobalHashGroup(BufferManager &buffer_manager, const Orders &partitions, const Orders &orders,
 	                      const Types &payload_types, idx_t max_mem, bool external)
-	    : memory_per_thread(max_mem), count(0), partition_layout(partitions) {
+	    : memory_per_thread(max_mem), count(0) {
 
 		RowLayout payload_layout;
 		payload_layout.Initialize(payload_types);
 		global_sort = make_unique<GlobalSortState>(buffer_manager, orders, payload_layout);
 		global_sort->external = external;
+
+		partition_layout = global_sort->sort_layout.GetPrefixComparisonLayout(partitions.size());
 	}
 
 	void Combine(LocalSortState &local_sort) {
@@ -96859,7 +96989,8 @@ struct DateDiff {
 	struct WeekOperator {
 		template <class TA, class TB, class TR>
 		static inline TR Operation(TA startdate, TB enddate) {
-			return Date::Epoch(enddate) / Interval::SECS_PER_WEEK - Date::Epoch(startdate) / Interval::SECS_PER_WEEK;
+			return Date::Epoch(Date::GetMondayOfCurrentWeek(enddate)) / Interval::SECS_PER_WEEK -
+			       Date::Epoch(Date::GetMondayOfCurrentWeek(startdate)) / Interval::SECS_PER_WEEK;
 		}
 	};
 
@@ -104318,18 +104449,12 @@ void SinkDataChunk(Vector *child_vector, SelectionVector &sel, idx_t offset_list
 static void ListSortFunction(DataChunk &args, ExpressionState &state, Vector &result) {
 	D_ASSERT(args.ColumnCount() >= 1 && args.ColumnCount() <= 3);
 	auto count = args.size();
-	Vector &lists = args.data[0];
+	Vector &input_lists = args.data[0];
 
 	result.SetVectorType(VectorType::FLAT_VECTOR);
 	auto &result_validity = FlatVector::Validity(result);
 
-	for (auto &v : args.data) {
-		if (v.GetVectorType() != VectorType::FLAT_VECTOR && v.GetVectorType() != VectorType::CONSTANT_VECTOR) {
-			v.Flatten(count);
-		}
-	}
-
-	if (lists.GetType().id() == LogicalTypeId::SQLNULL) {
+	if (input_lists.GetType().id() == LogicalTypeId::SQLNULL) {
 		result_validity.SetInvalid(0);
 		return;
 	}
@@ -104344,15 +104469,18 @@ static void ListSortFunction(DataChunk &args, ExpressionState &state, Vector &re
 	LocalSortState local_sort_state;
 	local_sort_state.Initialize(global_sort_state, buffer_manager);
 
+	// this ensures that we do not change the order of the entries in the input chunk
+	VectorOperations::Copy(input_lists, result, count, 0, 0);
+
 	// get the child vector
-	auto lists_size = ListVector::GetListSize(lists);
-	auto &child_vector = ListVector::GetEntry(lists);
+	auto lists_size = ListVector::GetListSize(result);
+	auto &child_vector = ListVector::GetEntry(result);
 	UnifiedVectorFormat child_data;
 	child_vector.ToUnifiedFormat(lists_size, child_data);
 
 	// get the lists data
 	UnifiedVectorFormat lists_data;
-	lists.ToUnifiedFormat(count, lists_data);
+	result.ToUnifiedFormat(count, lists_data);
 	auto list_entries = (list_entry_t *)lists_data.data;
 
 	// create the lists_indices vector, this contains an element for each list's entry,
@@ -104449,8 +104577,6 @@ static void ListSortFunction(DataChunk &args, ExpressionState &state, Vector &re
 		child_vector.Flatten(sel_sorted_idx);
 	}
 
-	result.Reference(lists);
-
 	if (args.AllConstant()) {
 		result.SetVectorType(VectorType::CONSTANT_VECTOR);
 	}
@@ -141402,7 +141528,7 @@ Value ForceCompressionSetting::GetSetting(ClientContext &context) {
 //===--------------------------------------------------------------------===//
 void HomeDirectorySetting::SetLocal(ClientContext &context, const Value &input) {
 	auto &config = ClientConfig::GetConfig(context);
-	config.home_directory = input.IsNull() ? input.ToString() : string();
+	config.home_directory = input.IsNull() ? string() : input.ToString();
 }
 
 Value HomeDirectorySetting::GetSetting(ClientContext &context) {
@@ -151925,6 +152051,7 @@ unique_ptr<NodeStatistics> StatisticsPropagator::PropagateStatistics(LogicalGet
 
 
 
+
 namespace duckdb {
 
 void StatisticsPropagator::PropagateStatistics(LogicalComparisonJoin &join, unique_ptr<LogicalOperator> *node_ptr) {
@@ -151954,10 +152081,15 @@ void StatisticsPropagator::PropagateStatistics(LogicalComparisonJoin &join, uniq
 					// semi or inner join on false; entire node can be pruned
 					ReplaceWithEmptyResult(*node_ptr);
 					return;
-				case JoinType::ANTI:
-					// anti join: replace entire join with LHS
-					*node_ptr = move(join.children[0]);
+				case JoinType::ANTI: {
+					// when the right child has data, return the left child
+					// when the right child has no data, return an empty set
+					auto limit = make_unique<LogicalLimit>(1, 0, nullptr, nullptr);
+					limit->AddChild(move(join.children[1]));
+					auto cross_product = LogicalCrossProduct::Create(move(join.children[0]), move(limit));
+					*node_ptr = move(cross_product);
 					return;
+				}
 				case JoinType::LEFT:
 					// anti/left outer join: replace right side with empty node
 					ReplaceWithEmptyResult(join.children[1]);
@@ -151985,10 +152117,15 @@ void StatisticsPropagator::PropagateStatistics(LogicalComparisonJoin &join, uniq
 				} else {
 					// this is the only condition and it is always true: all conditions are true
 					switch (join.join_type) {
-					case JoinType::SEMI:
-						// semi join on true: replace entire join with LHS
-						*node_ptr = move(join.children[0]);
+					case JoinType::SEMI: {
+						// when the right child has data, return the left child
+						// when the right child has no data, return an empty set
+						auto limit = make_unique<LogicalLimit>(1, 0, nullptr, nullptr);
+						limit->AddChild(move(join.children[1]));
+						auto cross_product = LogicalCrossProduct::Create(move(join.children[0]), move(limit));
+						*node_ptr = move(cross_product);
 						return;
+					}
 					case JoinType::INNER:
 					case JoinType::LEFT:
 					case JoinType::RIGHT:
@@ -152105,6 +152242,7 @@ unique_ptr<NodeStatistics> StatisticsPropagator::PropagateStatistics(LogicalJoin
 	// then propagate into the join conditions
 	switch (join.type) {
 	case LogicalOperatorType::LOGICAL_COMPARISON_JOIN:
+	case LogicalOperatorType::LOGICAL_DELIM_JOIN:
 		PropagateStatistics((LogicalComparisonJoin &)join, node_ptr);
 		break;
 	case LogicalOperatorType::LOGICAL_ANY_JOIN:
@@ -171765,7 +171903,7 @@ LogicalType Transformer::TransformTypeName(duckdb_libpgquery::PGTypeName *type_n
 
 		result_type = LogicalType::MAP(move(children));
 	} else {
-		int8_t width, scale;
+		int64_t width, scale;
 		if (base_type == LogicalTypeId::DECIMAL) {
 			// default decimal width/scale
 			width = 18;
@@ -180863,6 +181001,9 @@ BoundStatement Binder::BindSummarize(ShowStatement &stmt) {
 
 
 
+
+
+
 //===----------------------------------------------------------------------===//
 //                         DuckDB
 //
@@ -180897,10 +181038,6 @@ protected:
 
 
 
-
-
-
-
 //===----------------------------------------------------------------------===//
 //                         DuckDB
 //
@@ -180933,6 +181070,8 @@ public:
 };
 } // namespace duckdb
 
+
+
 #include <algorithm>
 
 namespace duckdb {
@@ -181103,10 +181242,10 @@ BoundStatement Binder::Bind(UpdateStatement &stmt) {
 		if (column.Generated()) {
 			throw BinderException("Cant update column \"%s\" because it is a generated column!", column.Name());
 		}
-		if (std::find(update->columns.begin(), update->columns.end(), column.Oid()) != update->columns.end()) {
+		if (std::find(update->columns.begin(), update->columns.end(), column.StorageOid()) != update->columns.end()) {
 			throw BinderException("Multiple assignments to same column \"%s\"", colname);
 		}
-		update->columns.push_back(column.Oid());
+		update->columns.push_back(column.StorageOid());
 
 		if (expr->type == ExpressionType::VALUE_DEFAULT) {
 			update->expressions.push_back(make_unique<BoundDefaultExpression>(column.Type()));
@@ -181188,7 +181327,20 @@ BoundStatement Binder::Bind(VacuumStatement &stmt) {
 			auto &get = (LogicalGet &)*ref->get;
 			columns.insert(columns.end(), get.names.begin(), get.names.end());
 		}
+
+		case_insensitive_set_t column_name_set;
+		vector<string> non_generated_column_names;
 		for (auto &col_name : columns) {
+			if (column_name_set.count(col_name) > 0) {
+				throw BinderException("Vacuum the same column twice(same name in column name list)");
+			}
+			column_name_set.insert(col_name);
+			auto &col = ref->table->GetColumn(col_name);
+			// ignore generated column
+			if (col.Generated()) {
+				continue;
+			}
+			non_generated_column_names.push_back(col_name);
 			ColumnRefExpression colref(col_name, ref->table->name);
 			auto result = bind_context.BindColumn(colref, 0);
 			if (result.HasError()) {
@@ -181196,17 +181348,29 @@ BoundStatement Binder::Bind(VacuumStatement &stmt) {
 			}
 			select_list.push_back(move(result.expression));
 		}
-		auto table_scan = CreatePlan(*ref);
-		D_ASSERT(table_scan->type == LogicalOperatorType::LOGICAL_GET);
-		auto &get = (LogicalGet &)*table_scan;
-		for (idx_t i = 0; i < get.column_ids.size(); i++) {
-			stmt.info->column_id_map[i] = get.column_ids[i];
-		}
+		stmt.info->columns = move(non_generated_column_names);
+		if (!select_list.empty()) {
+			auto table_scan = CreatePlan(*ref);
+			D_ASSERT(table_scan->type == LogicalOperatorType::LOGICAL_GET);
 
-		auto projection = make_unique<LogicalProjection>(GenerateTableIndex(), move(select_list));
-		projection->children.push_back(move(table_scan));
+			auto &get = (LogicalGet &)*table_scan;
+
+			D_ASSERT(select_list.size() == get.column_ids.size());
+			D_ASSERT(stmt.info->columns.size() == get.column_ids.size());
+			for (idx_t i = 0; i < get.column_ids.size(); i++) {
+				stmt.info->column_id_map[i] = ref->table->columns[get.column_ids[i]].StorageOid();
+			}
+
+			auto projection = make_unique<LogicalProjection>(GenerateTableIndex(), move(select_list));
+			projection->children.push_back(move(table_scan));
 
-		root = move(projection);
+			root = move(projection);
+		} else {
+			// eg. CREATE TABLE test (x AS (1));
+			//     ANALYZE test;
+			// Make it not a SINK so it doesn't have to do anything
+			stmt.info->has_table = false;
+		}
 	}
 	auto vacuum = make_unique<LogicalSimple>(LogicalOperatorType::LOGICAL_VACUUM, move(stmt.info));
 	if (root) {

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 "9656abd17"
-#define DUCKDB_VERSION "v0.5.1-dev7"
+#define DUCKDB_SOURCE_ID "62c4acd27"
+#define DUCKDB_VERSION "v0.5.1-dev72"
 //===----------------------------------------------------------------------===//
 //                         DuckDB
 //
@@ -27469,7 +27469,8 @@ public:
 public:
 	template <class T, class BASE>
 	static string ToString(const T &entry) {
-		return entry.left->ToString() + " " + ExpressionTypeToOperator(entry.type) + " " + entry.right->ToString();
+		return StringUtil::Format("(%s) %s (%s)", entry.left->ToString(), ExpressionTypeToOperator(entry.type),
+		                          entry.right->ToString());
 	}
 };
 } // namespace duckdb