My understanding so far( anyone can correct me if i'm wrong)
Answer 1: When you publish the workbook with the extracted data, the actual datasource is still on your local machine and the workbook just makes a reference of the location of the datasource in the Postgres. The actually datasource is not pushed to the server( when using xls/csv's etc) but their connection is published( datasource references)
Answer 2: when you publish just the datasource and schedule them, they get refreshed independently of the workbook. Workbook has no reference these sources and therefore, your workbook will not be refreshed.
- Scheduling the workbook will still look for the datasource that was referenced when originally created.
NOTE: Once you published both the sources and the workbook independently, if you replace the workbook datasource to the one of the server, now they're connected and updating the datasource will update the workbook.
- Scheduling a workbook will ONLY refresh your workbook and no other workbook will be refreshed though they're built on same DS.
- Scheduling a datasource will refresh all the workbooks associated. This way you can avoid the multiple refreshes on the same datasource. You can create a DS with all calculations/definitions and publish to server and anyone else using same source can connect to this DS located on the server.
Hope this helps,
How do you associate a workbook on the server to the datasource???
If I understand this, you connected to the datasources and created the workbook in the same desktop session. You then published the datasources as extracts and also published the workbook.
If that is the case, then my comment is this:
A common issue where I work was for someone to connect to a SQL server, create an extract and publish it to Tableau server and then immediately start designing a workbook. The key here is that they were still connected to the SQL server and never actually made the connection to the extract prior to designing the workbook. So when the extract was refreshed, the workbook never was. The workbook was also slow because it was doing a live connection back to SQL. It didn't matter that Use Extract was selected, it was still a connection back to SQL. (At least that's what I noticed.)
What should have been done was to close the SQL connection after publishing the extract. Then after doing so, opening a connection to Tableau server and selecting the newly created extract. Only THEN should they start designing the workbook. This ensures the workbook is directly connected to the extract.
Hope that helps.