R script visual not working in Service

[luxpbi]

Hi, 

I've created a R visual that load data from RData Files. Everything is working fine in desktop. 

But I have problem when I upload it to the cloud, the script doesn't found the path of the files because they are in local ... obiously. 

How can I solve this?

Thanks in advance!

[Stachu]

I think you need data gateway so the service can get access to your local files
https://docs.microsoft.com/en-us/power-bi/refresh-desktop-file-local-drive

[luxpbi]

Hi @Stachu,

The problem is just with the visual. 

I attach photos in case it helps to see the problem.

[TomMartens]

Hey,

as far as I understand your R script that works perfectly in your Power BI Desktop contains basically these 3 steps:

• read source data
• prepare the data
• create the visual

these 3 steps in one single r script will wonk in PBI Desktop.

But this will not work in the service, the reason is:
R script visuals or Custom Visuals (R based) are running in a kind of sandbox and are not allowed to access "external" ressources.

For this reason you have to split your R scipt.

• reading the source data (this can be done by using the R script source connection)
  this step has to return a table object to the Power BI data model
  be aware that it is necessary to install the on-premises data gateway in personal mode
• create the r visualbased on the Power BI table

The data preparation (if necessary) can either be done in the first step or also in the r script that creates the visual

This link will maybe also useful:

https://docs.microsoft.com/en-us/power-bi/desktop-r-in-query-editor

Regards,

Tom

[luxpbi]

Hi @TomMartens,

First of all, thank you for your answer. 

If I put the data in my Power BI model with an R script as data source, I have a limitation in the R visual script of 150.000 rows of data isn't it?

The fact that my data is out of Power BI and I it's in my R script is to avoid that limitation.

That's the way my R script works. 

• Load data 
• Prepare data 
• Pass parameters with Power BI (the idea is that the script will change with the inputs of the users)
• Create the visual

As I said in Power BI Desktop everything is working fine. The problem ocurs in the service, I understand the sandbox you are talking about, but is the a way to solve my situation?

Thanks again.

Regards.

[TomMartens]

Hey,

okay now it becomes clearer ...

Unfortunately there is this data point limit of 150000.

If you can't "compress" the number of data points there will be no solution.

Once I had been challenged with a similar problem ...

Lets assume your dataset looks like this

A | this | 1

A | this | 1

A | that | 2

I prepared the dataset that I passed to the R script visual in this way (I collapsed the dataset)

A | this | 1 | 2

A | that | 2 | 1

Inside the R script visual I then "expanded" the dataset again. I used the R package data.table for the collapsing and also for the expansion.

if you have to "visualize" more than 150000 datapoints at the same time, also considering filtering the dataset / the dataframe by Power BI Slicer and you are not able to collapse the dataset/dataframe that is passed to the R script visual, there is no solution - unfortunately.

Regards,

Tom

[luxpbi]

Hi Tom, 

Thank you very much for your answer. 

So I think for the moment the only solution in to make it work in Power BI Desktop.

Do you know if you can manage permisions in the desktop ? I don't want user to be able to manipulate data for example...

Regards.

[Anonymous]

Hey @luxpbi

I'm jumping over to this thread as I saw you opened a new question specific to getting this working in the service. Tom's idea to collapse and expand is really neat. I also wanted to ask, is your R visual performing any aggregations that you could perform in your data model before importing into the visual? 

Thanks,

Parker