[fix](csv-reader) fix column split error when there is escape charact…

…er (#34364) (#34505)
apache · May 8, 2024 · 0a38945 · 0a38945
1 parent 21db960
commit 0a38945
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/be/src/vec/exec/format/file_reader/new_plain_text_line_reader.cpp b/be/src/vec/exec/format/file_reader/new_plain_text_line_reader.cpp
@@ -143,13 +143,12 @@ void EncloseCsvLineReaderContext::_on_normal(const uint8_t* start, size_t& len)
 }
 
 void EncloseCsvLineReaderContext::_on_pre_match_enclose(const uint8_t* start, size_t& len) {
-    bool should_escape = false;
     do {
         do {
             if (start[_idx] == _escape) [[unlikely]] {
-                should_escape = !should_escape;
-            } else if (should_escape) [[unlikely]] {
-                should_escape = false;
+                _should_escape = !_should_escape;
+            } else if (_should_escape) [[unlikely]] {
+                _should_escape = false;
             } else if (start[_idx] == _enclose) [[unlikely]] {
                 _state.forward_to(ReaderState::MATCH_ENCLOSE);
                 ++_idx;

diff --git a/be/src/vec/exec/format/file_reader/new_plain_text_line_reader.h b/be/src/vec/exec/format/file_reader/new_plain_text_line_reader.h
@@ -135,6 +135,7 @@ class EncloseCsvLineReaderContext final
 
     inline void refresh_impl() {
         _idx = 0;
+        _should_escape = false;
         _result = nullptr;
         _column_sep_positions.clear();
         _state.reset();
@@ -168,6 +169,7 @@ class EncloseCsvLineReaderContext final
     const size_t _column_sep_len;
 
     size_t _idx = 0;
+    bool _should_escape = false;
 
     const std::string _column_sep;
     std::vector<size_t> _column_sep_positions;