Using PowerAutomate to refresh a Dataset

[cbschley]

Has anyone used PowerAutomate to refresh a dataset after more than 1 Dataflow is refreshed? The UI seems straightforward if you're only checking one Dataflow, but it doesn't seem to let you check more than 1 Dataflow to trigger the dataset refresh.

So that has me thinking about how we set up our Dataflows to begin with. Essentially we re-use dataflows in multiple dashboards, so it made sense to only load them once, and call those Dataflows in the appropriate dashboard. However now that I'm looking into PowerAutomate, I may need to load everything I need for one dashboard into a single Dataflow, so that I can automate the dataset refreshes.

Curious to know if anyone has run into similar issues, and how you handled it. Thanks!

[lbendlin]

In Power Automate you can create processes where a dataflow refresh would increase a counter (for example). Then you can have a separate flow that checks if that counter has reached 4  (for example), and trigger the dataset refresh accordingly.

However - as you are considering - the beauty of dataflow refreshes is that they can be done as needed rather than on a common schedule.  So you can decide if a  dataflow refresh does warrant an immediate dataset refresh or not.

[cbschley]

Is it possible to have that separate flow where you check counters be automated? I like the idea of that solution, just not sure how to execute on it. I don't see an option in automated cloud flows to monitor a variable or counter like you described.

[lbendlin]

You need a place to store that counter.  Easiest would be a file in a sharepoint document library, or a cell in an Excel file.  A glorified semaphore.

[cbschley]

So lets say I'm using Excel, which is what I'm trying in this case. How do I set up an automated cloud flow monitor that excel file? 

I currently have the flow set up to add new rows into an Excel document with all of the dataflow information, as well as a variable set to "1" if it was a successful flow. I don't understand what the next step would be here. 

[lbendlin · Accepted solution]

let's assume you have four dataflows. You have an excel file where each dataflow adds a row once finished. Then you have a separate flow that runs periodically (or on a file change trigger) and checks if that table has four rows. If yes - initiate the dataset refresh.  

[cbschley]

Ok that worked, using the file-change trigger. There were some nuances with the variables converting from string to float, but otherwise was able to get it to run successfully. Also added an action to clear out the excel file after successful completion so that it starts fresh each time it succeeds.