R/W endpoints

[Anonymous]

Hi all,
Since introducing of R/W endpoints i was super excited and i hardly waited to use some of the tools out there to be able to handle my datasets while they are online. 

Some of my datasets on dedicated (premium) capacity are really huge and downloading and uploading 1GB dataset just to do a simple change is not a job you would usually enyoj doing, but if you are using Tabular editor and R/W endpoints this could be done in 15 seconds!!

I tried this multiple times, and as a result measures gets updated and works fine, change is super fast, but I am not able to dowload it as PBIX any more. I get following message when i try to donwload dataset or report linked to changed dataset "We couldn't export to .pbix format. Please try again later or contact support. If you contact support, please provide these details."

When I reached supprot they found follwoing exception on their side:

Microsoft.PowerBI.PublishRefreshService.ExportReportDisabledForModelModifiedByXmlaEndpointException

and they said it will be fixed on 7th of April, and now even after one month it is not working

Is there someone who can share their experience? Does this work for you? Is somehow my PBI "cluster" not fixed while other ones are working?

[GilbertQ]

Hi there

Once your PBIX goes into Premium and is using R/W you can no longer download the PBIX because it is all changed within the Power BI Service.

You are more likely to modify your existing PBIX and upload it again (The challenge is that you will have to reload all the data)

There are changes coming later to make incremental changes to the PBIX and then upload only the changes.

[Anonymous]

Hi @GilbertQ 

Do you know any article online on this subject? I was watching that webinar from Christian Wade few months ago, and he never talked about any restrictions.

In my case after changing dataset on service with tabular editor i cannot longer download both dataset and reports that are using live connection to this dataset. This does not make any sense, i already have ticket open for this but MS support is just confused and giving me generic answer ("hey we will check and let you know") when i ask for status. 

So far in weeks they did nothing to provide any solution.

I know it is har to solve this issues like this on community so i was writing this post more to get an idea does anyone use this feature without any problems?

[GilbertQ]

Hi there

What I do is I always have a local copy of the dataset where I model and make my changes.

Only once that is done do I then deploy the changes to the PBI dataset.

This is how I would suggest developing and is considered best practice because you always work offline with a local copy. And ensures that you do not deploy a change that could break the entire model.

[Anonymous]

Thanks @GilbertQ 

I did it like that until they intruduced R/W endpoints for dedicated capacity. By looking on their presentatison what could be done on model while dataset is on service i tought "hey this is really going to help me with those huge datasets".

But at the end it is just not working 😕, so i will try go old way with dowloading datasets and working with local copy. It will make me less productive with large datasets, but at the moment this is probably only way to be on the safe side.

[GilbertQ]

Hi there

You can certainly do that the challenge is that the dataset only then gets changed in the service.

There is a way to script out the dataset and create it locally.