I'm sorry to hear your frustrations, and it's certainly a concern we're aware of. It's important to point out that the pain you describe is largely centered on Data Server data sources which have extracts. We're also aware of other issues with managing extracts, such as the need to publish a data source but defer the act of creating an extract until it reaches Tableau Server. It's a testament to how useful our customers find the Data Engine that we hear such feedback for better management of Data Engine extracts.
We aim to introduce new features with solid integration with all other relevant features, but we rely on feedback like yours to help identify gaps and clarify the priority for completing and expanding the feature in future releases. I think it's disingenuous to call this approach a complete farce, but I understand if your frustration drove you to express yourself that way. I appreciate your cheerful support of Tableau elsewhere on the web, and I hope you are aware how much we always strive to please our product champions such as yourself.
Keep sending us your constructive criticism!
Hi Robert. My intent wasn't to infuriate, so I apologize if I did. Regardless, the term Data Editor is, I think we can both agree, misleading.
As for your comment "the need to publish a data source but defer the act of creating an extract until it reaches Tableau Server", you can do that. All you have to do is create a calculate field with NOW() as the formula, add that the the extract on desktop as a filter, choose the date/time, then the extract will have zero records. Then you can publish it to Server, it will be very small which is an added benefit, then schedule the extract to run on Server. Since NOW is now false, the extract will run just fine.
Back to my comments. My question wasn't answered. But also, Server will not remember blended calculations in the same workbook, though it says it does. On top of that, blended formulas get renamed with some strange "oracle.blahblahblah" name for the secondary source.
It's very frustrating to say the least.
1 of 1 people found this helpful
I'm familiar with that hack, but it's not an elegant solution and I hope we can address it in a future release.
As for your concerns, I definitely understand how frustrating this can be. The example you describe with data blending actually sounds like a bug, so I encourage you to report it to our support team.
Regarding your question, I assumed it was rhetorical but I'll try to give some background here that may provide the context you were looking for. The purpose of the data source editor is to provide Desktop users with the means of extending a Server data source by layering new fields into the data model, such as calculations, ad-hoc groups, sets, etc. When a user connects to a Server data source, they are not working with the data source directly but instead are performing operations through the Data Server running on Tableau Server. The local data source in Desktop is what supports layering the data model as I described above, and this is the data source which you are working with in the "data source editor". You are correct that you must download the original data source from Tableau Server if you wish to work with it directly and make edits in Desktop.
Does this help?
Yes, that helps a lot. I doesn't make things better, but I understand how it works better now.
It'd be great if you could edit a Server data source directly and republish it. It seems like unnecessary, often time-consuming gymnastics to have to go through to simply change a calculation if it's wrong. I go from less than 5 seconds of work to over 30 minutes in some cases.
It's incredibly frustrating.
Thanks for the explanation Robert. See you at TCC.