vincbeck commented on code in PR #22737:
URL: https://github.com/apache/airflow/pull/22737#discussion_r844051248
##########
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:
> What will happen if I check for 300 keys? If I get it right this will make
300 API calls per poke? If so I think it's not good.
I agree but I also need to be backward compatible. That's the current
behavior in `S3PrefixSensor` so I cannot really change this logic. I cannot
either use some list APIs because the scope of these APIs are per bucket, and I
can imagine providing 300 different buckets
> Also if I poke for bucket_key=['a','b'] and 'a' is present but 'b' is not
why should I keep poke for a?
That's a good point. I can look into a way to optimize that
--
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]
