You should look at OneDrive publishing rather than push datasets for that. However, probably you have to schedule a refresh.

It is not related to the data source, it is up to you to build the technique to push updates, though.

You should use an automation process to run powershell, on Azure or on a PC - you can also look for Power Automate.

The command-line app is for testing purposes only. In order to push data into the dataset, you need to call the Power BI REST API through a service in the cloud that intercepts new data in your system. As a polling technique and no-code solution, you can use Power Automate to do the job.

Read the article, you should install the PowerShell cmdlet first.

Push dataset have the unique feature of updating the dashboard tiles as soon as new data arrive. Aggregations in DirectQuery should be refreshed periodically for the entire table (no partitioning) so currently there is no way to have an hybrid tables (part DirectQuery and part Import) that would solve the issue. However, even with hybrid tables you would still have a scalability issue with the DirectQuery part in case you have many users.

@@SQLBI Thanks a lot for explanation

It's hard to implement real-time updates with that architecture: you should create a system (with Power Automate or a script/service) that polls the blob storage looking for new files very often. Not sure it would be much different than a frequent incremental update (but it would be every 1-2 hours, not every minute).

@@SQLBIThanks for the reply! Maybe it would make sense to add an additional layer to transform data into table. It would be a good idea to cover the implmentation of the library presented in this video for Azure Cloud DB connection

Unfortunately, RLS is not supported in push datasets.

Please, can you clarify the question?

Push datasets don't have the limitations of DirectQuery. Power BI matrix visual (which is the way you obtain a pivot transformation) are compatible - but that's only at visual level, if you are talking about M, then push datasets don't have anything like that, you must transform the data upfront.

The "real-time" behavior could be good for a monitor that shows the information updated real-time. We call them "vanity projects" as they do not have a real business value, but they could be a motivational object. We've seen them used in monitor in the hall of a company (or in private areas when there are sensitive company data!).

Perhaps when it gets to the level of investor meetings? Other than that, as said, vanity.

The entire idea behind these real-time dashboards is to implement "vanity projects" without creating a dedicated infrastructure that is way more expensive to maintain. The daily cleanup partially solves the issue you mentioned. Clearly, you must have an error tolerance about the live data.

​@@SQLBIi need your personal contact please. I have a project in mining industry

No you cannot add slicers to a dashboard.

You would not remove the scalability issue for DirectQuery and you would lose the automatic refresh of the dashboard tile when there is new data. Performance depends on the implementation.

Yes you can! Just pay attention to what you use in Power Automate to detect the presence of new rows, because a polling *might* be resource consuming for the server. But it definitely works!

You can either copy the model.bim file manually or connect Tabular Editor to the server and retrieve the model from the server. However, we suggest that you use the Tabular Editor forum for these questions: github.com/TabularEditor/TabularEditor/issues

@@SQLBI Thanks a mill, I did it manually before will try how to retrieve it from the server. thanks again

Tabular Editor 2 is free and open source: www.sqlbi.com/tools/tabular-editor/ and github.com/TabularEditor/TabularEditor

Push datasets update dashboards as soon as new data comes in. Dual mode can keep data in Power Query, but you always have to trigger a dashboard refresh in order to see new data.

Please, write the technical questions in the comments of the related article and make sure you follow all the steps described here: www.sqlbi.com/articles/implementing-real-time-updates-in-power-bi-using-push-datasets-instead-of-directquery/

It depends. You cannot have automatic updates as shown in the dashboard, you could implement automatic refresh in the report. However, every report sends query to the DirectQuery partition, potentially introducing DirectQuery latency to every report (where is your data source?) and creating a bigger workload on the datasource (read: consider scalability for your data source). The hybrid tables should be easier to implement if the data source fits well in the DirectQuery scenario and you are able to configure the partition wisely. The push datasets are a good alternative for data sources that would not perform well in DirectQuery and/or you have transformation that don't comply with the folding requirements for hybrid tables. As usual, every tool has pros and cons, having more options is better because you have more choices and depending on the scenario you can choose the best one.

You can find it here: www.sqlbi.com/blog/alberto/2021/02/14/on-my-recording-gear/

No, push datasets only works in Power BI service.

What do you mean?

@@SQLBI I mean that works really well when we are increasing the database, but seems that when we need to update some figures don't work, is that right?

For the push dataset part, an update can be obtained by pushing a compensating transaction. The biggest issue is that you have to identify the change first. Once you know that a transaction changed from $10 to $8, you have to push a compensating transaction of $-2 Or, you can just ignore that, knowing that these adjustments will be reflected in the next daily refresh. For example, in an implementation we made, we just ignore refunds made in the day, because they will appear the following day in the regular daily refresh.