ephraimbuddy commented on code in PR #60864:
URL: https://github.com/apache/airflow/pull/60864#discussion_r2726532388
##########
airflow-core/src/airflow/dag_processing/manager.py:
##########
@@ -989,8 +990,28 @@ def _add_new_files_to_queue(self, known_files: dict[str,
set[DagFileInfo]]):
def _resort_file_queue(self):
if self._file_parsing_sort_mode == "modified_time" and
self._file_queue:
- files, _ = self._sort_by_mtime(self._file_queue)
- self._file_queue = deque(files)
+ files_with_mtime: dict[DagFileInfo, float] = {}
+ mtime_changed = False
+
+ for file in list(self._file_queue):
+ try:
+ mtime = os.path.getmtime(file.absolute_path)
+ files_with_mtime[file] = mtime
+ stat = self._file_stats[file] # Creates entry via
defaultdict if missing
+ if stat.last_mtime != mtime:
+ mtime_changed = True
+ stat.last_mtime = mtime
+ except FileNotFoundError:
+ self.log.warning("Skipping processing of missing file:
%s", file)
+ self._file_stats.pop(file, None)
+ mtime_changed = True # Queue structure changed
+
+ if not mtime_changed:
+ return # No changes, skip sorting
+
+ # Sort by mtime descending and rebuild queue
+ sorted_files = [f for f, _ in sorted(files_with_mtime.items(),
key=itemgetter(1), reverse=True)]
Review Comment:
Nope. This is replicating what `_resort_by_mtime` does but optimizing by
avoiding unnecessory resorting.
New files would have most recent mtimes which is higher thus processed first
since it's by descending order. Older ones will be done last
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
