CODEWITHSUNDEEP
powerbi
Sauron
Sauron

Reputation: 6657

How to edit Large Power BI Dataset after TMSL refresh

So currently we have a smaller dataset saved off in SharePoint that is downloaded, make your changes, then publish to power bi service. Since the dataset that was edited had like no data in it, it took over 5 hours to load in power bi service, so had to go through TMSL. But now that it ran in TMSL (and being a large model) it only allows me to download a copy and connect to the dataset live:

enter image description here

which then opening in power bi desktop, does not allow me to edit the dataset at all and does not permit editing the parameters, etc.

enter image description here

enter image description here

Is there a way to actually edit this power bi dataset without having to go to the saved off sharepoint doc now (that would have to run through TMSL again for each change since the dataset has to 'catch up'?

(I remeber in the old Azure Analysis Services days you could edit the metadata, relationships, measures calculated columns in Visual Studio, then deploy the model, not needing to catch up on years worth of data, etc...)

Upvotes: 0

Views: 294

Answers (2)

trenton-ftw
trenton-ftw

Reputation: 972

You should look into Tabular Editor. This tool has a free, open source version (Tabular Editor 2) that solves many common scaling issues you will run into with maintaining larger datasets within Power BI Desktop.

It allows you to connect directly to a deployed model and edit the (TOM) model that is deployed without requiring any interaction with the data stored in the model. It is also extremely helpful in version control, deployment, and maintenance of models.

Upvotes: 1

pauliec
pauliec

Reputation: 451

When you connect to a dataset via live connection mode, you can't edit the dataset, parameters or data model within pbi desktop but there are a few things you can do to not have to go through the sp document and tmsl process

  1. My personal preference is to use Dataflows. You can go to your data, then add new and add the sp data as a dataflow. It makes it easily reusable to transformation logic and you can use it across different datasets and reports. Especially useful if you and or this report isn't the only consumer of the data.

      [fantastic video from Radacad][1]

  2. You could also use Sql Server Analysis Services (SSAS) or Azure Analysis Services. You could develop the data model in VS just like in days of yore and deploy the model to Analysis Services then connect Power BI to Analysis Services model using a live connection. Then you'd be able to change your data model without having to refresh the whole dataset in pbi.

  3. Not as much of a fan because I don't think it'll exactly fit the bill, but you could first try making some changes in the pbi service. E.g new measures, apply themes, modify report level filters, etc. but it won't permit you to modify the data model, relationships or parameters.

  4. While in development, using a smaller dataset can be a lifesaver. Then you can go back and change as needed on a test set before using the production dataset.

Upvotes: 0

Related Questions