I am working to publish and extract a data source to be used in our production workbooks. This data source is made up of 10 tables hosted in BigQuery in a sandbox dataset. 8 of these tables have just over 9 million rows, one of the tables has around 150 million rows, and one table is extremely small that is being used for custom groupings of the data.
I created Data Source 1 (DS1) using these 10 sandbox tables and was able to successfully publish and extract this data source. It is working fine and I can switch back to live connection and re-do the extract as needed (have done so at least once).
I then tried to recreate the data source using the same tables but in a production dataset instead, which then failed to extract in under the 2 hour timeout limit.
As I was curious why the production version failed but the sandbox version didn't, I re-created the sandbox data source from scratch, matching cardinality and join performance optimizations and such from DS1, into a new data source that is identical. Let's call this DS2.
While DS2 is the same as DS1, using the same tables and connections, and published into the same folder on Tableau, it is unable to create the extract in under 2 hours.
What I'm looking for is any potential causes that may result in this situation and any workarounds the community has used to get around this. I have tried extracting at different times during the day/night with no luck there either.
Note: I've been unsuccessful in extracting on desktop as well, but still open to solutions here too.
The tables are the exact same size in both sandbox and production datasets. One table is about 6GB, one is ~19GB, and the rest are <400MB each.
