Using Tableau Output tool with the Update Option

[craigja]

Hi,

Another team in my business is having an issue as they are trying to upload around 6 million rows of data from Alteryx to Tableau Online every day (and its taking around 50 minutes!).  We have asked why they upload 6 million records daily and its because some of them are backdated records going back years.  I think this might be due to something like a client changing their name and they then change all records of that client to the new name,  Anyway, what we are thinking is that they should only upload the updated records but I'm not sure how that works - we can obviously identify what records should be updated, but is it as simple as choosing the Update option and then only passing those records to the Tableau Output tool with the Update/Upsert option selected? 

[patrick_mcauliffe · Accepted solution]

Think of Tableau Data Extracts (or the Hyper files) that Tableau uses on the backend more like Excel files than like an indexed database.

You're pretty much stuck with a full overwrite every time (if you want to update).  You're better off setting up a SQL instance in the cloud and having Tableau Online connect to that.

[craigja]

Thanks for that - that was what prompted my original question, I could not figure out (as somebody who used to be a SQL dev) how with no index an update would work from Alteryx to Tableau.

We have suggested to the team in question that they might want to change their process and have 2 Tableau datasources - with a linking field (The client ID) and then they only update the Client Datasource when the client name changes, this saves them updating millions of old trades going back over a decade!