vincbeck commented on code in PR #22737:
URL: https://github.com/apache/airflow/pull/22737#discussion_r844186593
##########
airflow/providers/amazon/aws/sensors/s3.py:
##########
@@ -78,27 +80,32 @@ def __init__(
):
super().__init__(**kwargs)
self.bucket_name = bucket_name
- self.bucket_key = bucket_key
+ self.bucket_key = [bucket_key] if isinstance(bucket_key, str) else
bucket_key
self.wildcard_match = wildcard_match
self.aws_conn_id = aws_conn_id
self.verify = verify
self.hook: Optional[S3Hook] = None
- def _resolve_bucket_and_key(self):
+ def _resolve_bucket_and_key(self, key):
"""If key is URI, parse bucket"""
if self.bucket_name is None:
- self.bucket_name, self.bucket_key =
S3Hook.parse_s3_url(self.bucket_key)
+ return S3Hook.parse_s3_url(key)
else:
- parsed_url = urlparse(self.bucket_key)
+ parsed_url = urlparse(key)
if parsed_url.scheme != '' or parsed_url.netloc != '':
raise AirflowException('If bucket_name provided, bucket_key
must be relative path, not URI.')
+ return self.bucket_name, key
- def poke(self, context: 'Context'):
- self._resolve_bucket_and_key()
- self.log.info('Poking for key : s3://%s/%s', self.bucket_name,
self.bucket_key)
+ def _key_exists(self, key):
+ bucket_name, key = self._resolve_bucket_and_key(key)
+ self.log.info('Poking for key : s3://%s/%s', bucket_name, key)
if self.wildcard_match:
- return self.get_hook().check_for_wildcard_key(self.bucket_key,
self.bucket_name)
- return self.get_hook().check_for_key(self.bucket_key, self.bucket_name)
+ return self.get_hook().check_for_wildcard_key(key, bucket_name)
+
+ return self.get_hook().check_for_key(key, bucket_name)
+
+ def poke(self, context: 'Context'):
+ return all(self._key_exists(key) for key in self.bucket_key)
Review Comment:
The only way to optimize it would be to use `ListObjects` instead but there
are some caveats.
1. `ListObjects` works for a given bucket. So If you want check 300
different keys, all of them in different bucket, it would make the problem worse
2. It would make the code more complex because by using `ListObjects` you
need to handle the different pages, making the necessary checks that the file
is returned as part of `ListObjects`
3. By looking really quick on how current users use `S3PrefixSensor`
(https://github.com/search?q=S3PrefixSensor&type=code), most of it are single
file usage
Because of these points, in my opinion, it's not worth using `ListObjects`
instead of `HeadObject`
--
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]
