Power BI dataset refresh with Azure Data Factory

[Anonymous]

Hi Community,

Any help will be highly appreciated it. 

Thanks

[Anonymous]

@Anonymous Depency is a possibility. As i said before i am trying to reduce costs. Since i am loading to a embedded

server also here i am trying to reduce costs. My embedded server is down at night.

So i am loading my dwh at night and when my embedded server is up i am kicking off my PBI load

As far as Dependency goes there are several options.

The basic depency is if dataload fails do not refresh PBI. If you have seperate schedules you need something like a log table which you read prior to starting a refresh.

The more complex is this: Suppose you have 2 reports X and Z

X uses table a and table b

Z uses table c and table b

If table c fails you can refresh table X but not Z

If table b fails you cannot refresh table X and Z

In this case you need a different check.

So you could implement depency several ways. It all depends on what you want/need

[ibarrau]

Hi. It depends on the architecture. If you have some pipelines moving data from different sources to load a datawarehouse with your bi model. You are probably using data factory to orquestrate this. In that case it's a good idea that the last step in the data factory orquestration will refresh power bi. That way you keep consistency that everything will run after it finishes the last task. That way you prevent that the schedule is run and the data is not ready.

If you are using data factory in a different scenario to run some data but not all of them and you are not sure that after your process you can run a power bi refresh, then you can schedule it.

Hope it make sense for you.

[Anonymous]

There is no 'better'. Some of my reports are on schedule. Others i refresh using ADF.

Reason for the last is that my datasets sometimes are rather large and timeout while refreshing. When i upgrade the DTU of the SQL server the loading works fine. So i upgrade my SQL .. refresh the PBI dataset and then downgrade the SQL. This way i reduce the costs of the SQL server.