[Fix](parquet-reader) some fix for #5 (Fix and optimize parquet min-m…

…ax filtering).
morningman · Jul 24, 2024 · b75f5b8 · b75f5b8
1 parent 4eb4770
commit b75f5b8
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 10 deletions.
diff --git a/be/src/vec/exec/format/parquet/parquet_common.h b/be/src/vec/exec/format/parquet/parquet_common.h
@@ -437,16 +437,16 @@ class CorruptStatistics {
 private:
     static void warn_parse_error_once(const std::string& createdBy, const std::string_view& msg) {
         //if (!already_logged.exchange(true)) {
-            LOG(WARNING) << "Ignoring statistics because created_by could not be parsed (see "
-                            "PARQUET-251)."
-                            " CreatedBy: "
-                         << createdBy << ", msg: " << msg;
+        LOG(WARNING) << "Ignoring statistics because created_by could not be parsed (see "
+                        "PARQUET-251)."
+                        " CreatedBy: "
+                     << createdBy << ", msg: " << msg;
         //}
     }
 
     static void warn_once(const std::string_view& msg) {
         //if (!already_logged.exchange(true)) {
-            LOG(WARNING) << msg;
+        LOG(WARNING) << msg;
         //}
     }
 

diff --git a/be/src/vec/exec/format/parquet/parquet_pred_cmp.h b/be/src/vec/exec/format/parquet/parquet_pred_cmp.h
@@ -131,6 +131,8 @@ class ParquetPredicate {
 
         CppType min_value;
         CppType max_value;
+        std::unique_ptr<std::string> encoded_min_copy;
+        std::unique_ptr<std::string> encoded_max_copy;
         tparquet::Type::type physical_type = col_schema->physical_type;
         switch (col_val_range.type()) {
 #define DISPATCH(REINTERPRET_TYPE, PARQUET_TYPE)                           \
@@ -186,13 +188,13 @@ class ParquetPredicate {
         case TYPE_STRING:
             if constexpr (std::is_same_v<CppType, StringRef>) {
                 if (!use_min_max_value) {
-                    std::string min_copy(encoded_min);
-                    std::string max_copy(encoded_max);
-                    if (!_try_read_old_utf8_stats(min_copy, max_copy)) {
+                    encoded_min_copy = std::make_unique<std::string>(encoded_min);
+                    encoded_max_copy = std::make_unique<std::string>(encoded_max);
+                    if (!_try_read_old_utf8_stats(*encoded_min_copy, *encoded_max_copy)) {
                         return false;
                     }
-                    min_value = StringRef(min_copy);
-                    max_value = StringRef(max_copy);
+                    min_value = StringRef(*encoded_min_copy);
+                    max_value = StringRef(*encoded_max_copy);
                 } else {
                     min_value = StringRef(encoded_min);
                     max_value = StringRef(encoded_max);