I'm playing around with getting data extracts on our tableau server. Publishing them as a small view (so it is fairly quick) and refreshing them once the views are unrestricted. My current test is an extract based on combining 3 views in my database. (So I went into tableau, connect to data, choose my central view, then choose multiple tables and add in two more views. Each view is restricted to return one row to make sure publishing was easy. Save data connection and publish to server. Once publishing is complete, remove the restriction from each view so they return lots of rows. Then try a refresh on the server to get all the data in.)
Yesterday attempt one failed because someone broke something in the database, a function was recompiled, resulting in the view breaking... okay - no problem.
Attempt two (running overnight to avoid others breaking the database), failed because it took longer than 2 hours... Is there a setting to let it run longer?
Today we started the refresh again before coming up with a more clever idea to try and make the views work better. So I'm trying to improve the way the view runs and I'm trying to stop the running refresh so I can make some changes and start it again.
I investigated on this forum and found there is no way to stop the extract refresh other than restart the server... I'll put in/vote for an idea for that.
So I thought... why not break the view as happened yesterday, then the extract will fail... brute force option but less brute force than restarting the server. So I went and dropped the view that is central to my extract (- the one I chose first when creating the connection). However, much to my surprise, half an hour after I dropped the central view necessary for my extract, the refresh is still happily trundling along. How can this be?
So two questions:
1. Is there any way to allow a refresh to take longer than 2 hours on the server?
2. How does the data connection/refresh work if you have chosen multiple views.... Does it flatten the query? If so, how can dropping the central one not fail it? Does it load all data from one view and then add the extra columns from the other views? Does it take the SQL statements from the views and cache them while running the extract?
Thanks for any help/answers for both questions !