[ https://issues.apache.org/jira/browse/AIRFLOW-4226?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16937683#comment-16937683 ]
Kaxil Naik edited comment on AIRFLOW-4226 at 9/25/19 12:23 PM: --------------------------------------------------------------- Thanks [~jackjack10]. [~liblonga] I agree this should change and made consistent, allowing users to override project on task/operator level. I will assign this to myself and try to fix this soon. In the meantime, if anyone wants to work on it please let me know. was (Author: kaxilnaik): Thanks [~jackjack10]. [~liblonga] I agree this should change and made consistent, allowing users to override project on task/operator level. I will assign this to myself and try to fix this soon. > DataProcPySparkOperator gets project ID from gcp_conn_id rather than > project_id > ------------------------------------------------------------------------------- > > Key: AIRFLOW-4226 > URL: https://issues.apache.org/jira/browse/AIRFLOW-4226 > Project: Apache Airflow > Issue Type: Bug > Components: gcp > Reporter: Aaron Liblong > Priority: Minor > > DataProcPySparkOperator gets [the project ID for the cluster it > creates|https://github.com/apache/airflow/blob/master/airflow/contrib/operators/dataproc_operator.py#L1334] > from [the hook instantiated from > gcp_conn_id|https://github.com/apache/airflow/blob/master/airflow/contrib/operators/dataproc_operator.py#L1324] > rather than from its own project_id property. This is contrary to how every > other operator in this module works. Not sure if it's a bug or intended. -- This message was sent by Atlassian Jira (v8.3.4#803005)