uranusjr commented on code in PR #42343:
URL: https://github.com/apache/airflow/pull/42343#discussion_r1768169706


##########
airflow/models/taskinstance.py:
##########
@@ -2908,42 +2908,37 @@ def _register_dataset_changes(self, *, events: 
OutletEventAccessors, session: Se
                 for dataset_alias_event in events[obj].dataset_alias_events:
                     dataset_alias_name = 
dataset_alias_event["source_alias_name"]
                     dataset_uri = dataset_alias_event["dest_dataset_uri"]
-                    extra = dataset_alias_event["extra"]
-                    frozen_extra = frozenset(extra.items())
+                    frozen_extra = 
frozenset(dataset_alias_event["extra"].items())
+                    dataset_alias_names[(dataset_uri, 
frozen_extra)].add(dataset_alias_name)
 
-                    dataset_tuple_to_alias_names_mapping[(dataset_uri, 
frozen_extra)].add(dataset_alias_name)
+        class _DatasetModelCache(Dict[str, DatasetModel]):
+            log = self.log
 
-        dataset_objs_cache: dict[str, DatasetModel] = {}
-        for (uri, extra_items), alias_names in 
dataset_tuple_to_alias_names_mapping.items():
-            if uri not in dataset_objs_cache:
-                dataset_obj = 
session.scalar(select(DatasetModel).where(DatasetModel.uri == uri).limit(1))
-                dataset_objs_cache[uri] = dataset_obj
-            else:
-                dataset_obj = dataset_objs_cache[uri]
-
-            if not dataset_obj:
-                dataset_obj = DatasetModel(uri=uri)
-                dataset_manager.create_datasets(dataset_models=[dataset_obj], 
session=session)
-                self.log.warning("Created a new %r as it did not exist.", 
dataset_obj)
+            def __missing__(self, key: str) -> DatasetModel:
+                (dataset_obj,) = 
dataset_manager.create_datasets([Dataset(uri=key)], session=session)
                 session.flush()
-                dataset_objs_cache[uri] = dataset_obj
-
-            for alias in alias_names:
-                alias_obj = session.scalar(
-                    select(DatasetAliasModel).where(DatasetAliasModel.name == 
alias).limit(1)
-                )
-                dataset_obj.aliases.append(alias_obj)
+                self.log.warning("Created a new %r as it did not exist.", 
dataset_obj)
+                self[key] = dataset_obj
+                return dataset_obj
 
-            extra = {k: v for k, v in extra_items}
+        dataset_objs_cache = _DatasetModelCache(
+            (dataset_obj.uri, dataset_obj)
+            for dataset_obj in session.scalars(
+                select(DatasetModel).where(DatasetModel.uri.in_(uri for uri, _ 
in dataset_alias_names))
+            )
+        )
+        for (uri, extra_items), alias_names in dataset_alias_names.items():
+            dataset_obj = dataset_objs_cache[uri]
             self.log.info(
                 'Creating event for %r through aliases "%s"',
                 dataset_obj,
                 ", ".join(alias_names),
             )
             dataset_manager.register_dataset_change(
                 task_instance=self,
-                dataset=dataset_obj,
-                extra=extra,
+                dataset=dataset_obj.to_public(),

Review Comment:
   This was a bug—`dataset` here should be passed a Dataset, not DatasetModel. 
The `to_public` methods on DatasetModel was added just for this.



-- 
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]

Reply via email to