duckdb 0.5.1-dev271.0 → 0.5.1-dev282.0

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "duckdb",
   "main": "./lib/duckdb.js",
-  "version": "0.5.1-dev271.0",
+  "version": "0.5.1-dev282.0",
   "description": "DuckDB node.js API",
   "gypfile": true,
   "dependencies": {

package/src/duckdb.cpp

@@ -18941,7 +18941,8 @@ namespace duckdb {
 
 static unordered_map<column_t, string> GetKnownColumnValues(string &filename,
                                                             unordered_map<string, column_t> &column_map,
-                                                            bool filename_col, bool hive_partition_cols) {
+                                                            duckdb_re2::RE2 &compiled_regex, bool filename_col,
+                                                            bool hive_partition_cols) {
 	unordered_map<column_t, string> result;
 
 	if (filename_col) {
@@ -18952,7 +18953,7 @@ static unordered_map<column_t, string> GetKnownColumnValues(string &filename,
 	}
 
 	if (hive_partition_cols) {
-		auto partitions = HivePartitioning::Parse(filename);
+		auto partitions = HivePartitioning::Parse(filename, compiled_regex);
 		for (auto &partition : partitions) {
 			auto lookup_column_id = column_map.find(partition.first);
 			if (lookup_column_id != column_map.end()) {
@@ -18990,10 +18991,10 @@ static void ConvertKnownColRefToConstants(unique_ptr<Expression> &expr,
 // 	- s3://bucket/var1=value1/bla/bla/var2=value2
 //  - http(s)://domain(:port)/lala/kasdl/var1=value1/?not-a-var=not-a-value
 //  - folder/folder/folder/../var1=value1/etc/.//var2=value2
-std::map<string, string> HivePartitioning::Parse(string &filename) {
-	std::map<string, string> result;
+const string HivePartitioning::REGEX_STRING = "[\\/\\\\]([^\\/\\?\\\\]+)=([^\\/\\n\\?\\\\]+)";
 
-	string regex = "[\\/\\\\]([^\\/\\?\\\\]+)=([^\\/\\n\\?\\\\]+)";
+std::map<string, string> HivePartitioning::Parse(string &filename, duckdb_re2::RE2 &regex) {
+	std::map<string, string> result;
 	duckdb_re2::StringPiece input(filename); // Wrap a StringPiece around it
 
 	string var;
@@ -19004,6 +19005,11 @@ std::map<string, string> HivePartitioning::Parse(string &filename) {
 	return result;
 }
 
+std::map<string, string> HivePartitioning::Parse(string &filename) {
+	duckdb_re2::RE2 regex(REGEX_STRING);
+	return Parse(filename, regex);
+}
+
 // TODO: this can still be improved by removing the parts of filter expressions that are true for all remaining files.
 //		 currently, only expressions that cannot be evaluated during pushdown are removed.
 void HivePartitioning::ApplyFiltersToFileList(vector<string> &files, vector<unique_ptr<Expression>> &filters,
@@ -19011,6 +19017,7 @@ void HivePartitioning::ApplyFiltersToFileList(vector<string> &files, vector<uniq
                                               bool hive_enabled, bool filename_enabled) {
 	vector<string> pruned_files;
 	vector<unique_ptr<Expression>> pruned_filters;
+	duckdb_re2::RE2 regex(REGEX_STRING);
 
 	if ((!filename_enabled && !hive_enabled) || filters.empty()) {
 		return;
@@ -19019,7 +19026,7 @@ void HivePartitioning::ApplyFiltersToFileList(vector<string> &files, vector<uniq
 	for (idx_t i = 0; i < files.size(); i++) {
 		auto &file = files[i];
 		bool should_prune_file = false;
-		auto known_values = GetKnownColumnValues(file, column_map, filename_enabled, hive_enabled);
+		auto known_values = GetKnownColumnValues(file, column_map, regex, filename_enabled, hive_enabled);
 
 		FilterCombiner combiner;
 		for (auto &filter : filters) {

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 "705fc18d5"
-#define DUCKDB_VERSION "v0.5.1-dev271"
+#define DUCKDB_SOURCE_ID "3d23491fa"
+#define DUCKDB_VERSION "v0.5.1-dev282"
 //===----------------------------------------------------------------------===//
 //                         DuckDB
 //
@@ -26998,12 +26998,16 @@ class HivePartitioning {
 public:
 	//! Parse a filename that follows the hive partitioning scheme
 	DUCKDB_API static std::map<string, string> Parse(string &filename);
+	DUCKDB_API static std::map<string, string> Parse(string &filename, duckdb_re2::RE2 &regex);
 	//! Prunes a list of filenames based on a set of filters, can be used by TableFunctions in the
 	//! pushdown_complex_filter function to skip files with filename-based filters. Also removes the filters that always
 	//! evaluate to true.
 	DUCKDB_API static void ApplyFiltersToFileList(vector<string> &files, vector<unique_ptr<Expression>> &filters,
 	                                              unordered_map<string, column_t> &column_map, idx_t table_index,
 	                                              bool hive_enabled, bool filename_enabled);
+
+	//! Returns the compiled regex pattern to match hive partitions
+	DUCKDB_API static const string REGEX_STRING;
 };
 
 } // namespace duckdb