Shekharrajak opened a new pull request, #60016:
URL: https://github.com/apache/airflow/pull/60016
Fix clearing mapped task instances in same group to respect depth-first
execution (#40543)
## Problem
When clearing a specific mapped task instance within a mapped task group
(e.g., `etl.e[1]`), the system incorrectly cleared ALL downstream task
instances (`etl.t[]` and `etl.l[]`) instead of only the relevant ones with the
same `map_index` (`etl.t[1]` and `etl.l[1]`). This violated the depth-first
execution model where each mapped instance should operate independently.
## Solution
Added explicit handling in `_get_relevant_map_indexes()` to detect when two
tasks are siblings in the same mapped task group. When `ti_count ==
ancestor_ti_count`, the function now returns the same `map_index` directly,
ensuring depth-first execution is maintained.
## Changes
- **Modified** `airflow-core/src/airflow/models/taskinstance.py`:
- Added special case check in `_get_relevant_map_indexes()` for same-group
siblings
- Returns `map_index` directly when `ti_count == ancestor_ti_count`
- **Added tests** in `airflow-core/tests/unit/models/test_taskinstance.py`:
- `test_find_relevant_relatives_downstream_same_group_siblings`:
Parametrized test for map_indexes 0, 1, 2
- `test_find_relevant_relatives_upstream_same_group_siblings`: Test for
upstream direction
## Testing
All 11 `find_relevant_relatives` tests pass, including:
- 6 existing tests (still passing)
- 4 new tests validating the fix
--
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]
