You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@airflow.apache.org by "Flávio Clésio (JIRA)" <ji...@apache.org> on 2018/06/14 14:53:00 UTC
[jira] [Created] (AIRFLOW-2620) New property called
"RetainSameConnection"
Flávio Clésio created AIRFLOW-2620:
--------------------------------------
Summary: New property called "RetainSameConnection"
Key: AIRFLOW-2620
URL: https://issues.apache.org/jira/browse/AIRFLOW-2620
Project: Apache Airflow
Issue Type: Wish
Components: configuration, operators, redshift
Affects Versions: Airflow 2.0
Reporter: Flávio Clésio
Some ETL stuff relies on the -high- reusability of the connections in the middle of the data flow, and part of this is achieved using temporary tables.
I performed a test [1] in Airflow and the behaviour is the for each command that we use the "PostgresOperator" hits the database again, and if the DAG had a lot of connections it can be a nightmare.
As I said in SO, for several cases where you have cases of processing some tables in stages the reuse a temporary table across multiple tasks can be very useful and plus: can remove a lot of pressure in the RDBMS cause you won't need to hit several times.
That's why a purpose a feature of the connection called "RetainSameConnection" that will keep the same connection during the execution of the DAG.
Reference: [1] - [My question in Stack Overflow|https://stackoverflow.com/questions/50858770/airflow-retain-the-same-database-connection/] about it
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)