Hello, I'm writing a Beam pipeline that does some relatively expensive reads from BigQuery. I want to be able to run the pipeline in a development loop without racking up a huge bill.
I know BigQuery has support for query caching, but from the docs, that only works if you don't specify a destination table. For the purposes of development, I don't mind trading off stale data (i.e. reusing an existing destination table if it exists) to save money. Is there any way to do this now, or relevant any open issues? I did a quick pass through JIRA but couldn't find anything. Thanks, Matt