[airflow] branch main updated: Give the the test_xcom_ar_map test opportunity to flush the data (#33153)

[potiuk]

This is an automated email from the ASF dual-hosted git repository.

potiuk pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/airflow.git


The following commit(s) were added to refs/heads/main by this push:
     new 6b21b79f33 Give the the test_xcom_ar_map test opportunity to flush the 
data (#33153)
6b21b79f33 is described below

commit 6b21b79f33e245ff1612b1970d05ef692c41f15c
Author: Jarek Potiuk <ja...@potiuk.com>
AuthorDate: Sun Aug 6 16:12:56 2023 +0200

    Give the the test_xcom_ar_map test opportunity to flush the data (#33153)
    
    Using same session in different steps of the same test has the
    potential of not flishing/committing the changes between.
    
    Seems that the #33150 traded one flakiness with another. Attempting to
    make sure that the flash/commit is executed before the second run.
    
    Error:
    
    The test RuntimeError: number of values in row (0) differ from number of
    column processors (29)
    
    The error is strange however and indicates a bug in sqlite library.
---
 tests/models/test_xcom_arg_map.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tests/models/test_xcom_arg_map.py 
b/tests/models/test_xcom_arg_map.py
index 541ab6fcfa..76d52f4769 100644
--- a/tests/models/test_xcom_arg_map.py
+++ b/tests/models/test_xcom_arg_map.py
@@ -253,6 +253,9 @@ def test_xcom_map_nest(dag_maker, session):
     for ti in decision.schedulable_tis:
         ti.run(session=session)
 
+    session.flush()
+    session.commit()
+
     # Now "pull" should apply the mapping functions in order.
     decision = dr.task_instance_scheduling_decisions(session=session)
     for ti in decision.schedulable_tis: