Refreshing from multiple data sources, on different time of day

[danielmillion]

Hello,

My client's report is connected to 3 datasources :

1. SQL server #1

2. SQL server #2

3. CRM (Custome Query)

He asked me if data from CRM could refresh 3 times a day at xx:xx yy:yy zz:zz

and the data from SQL servers will refresh 8 times a day at thoes times + in between them.

I know i could do it by creating 2 datasets (SQL, CRM)

and then create a new report with two "power bi dataset" as dataset,

this way i could configure the refresh time he asked me to. 

My question is : is it considered best practice ? 

# I dont like that my client's sources are spread. (I have many clients and will have much more, and i dont want to manage a messy working environment, i like it when it is all in one place.)

# This method relies exclusivly on "Power BI dataset" as dataset which is a relativly new feature, but this method is a workaround, not a feature by itself.
# Requests can change. What if he will ask me to split again ? this will make me create the report from scratch completly.AGAIN.

# What about other clients with much much more complex systems that will require to manage seperatly 10 different datasources?

# I dont want to do something because i can, but even if i disagree, i will if it is not a bad practice.

* I do think this is an important anacdote to consider when working with new clients.

If there's any other way to achieve what i need i will be happy to hear.

Thank you dear community.

[collinq]

Hi @danielmillion ,

Since datasets are the level at which refreshes are scheduled the way you are suggesting is the way I have done it.  It does mean two datasets in this situation.  And yes, that means one report with two datasets but that is also common solution.

Your question is intriguing - if this is "best practice".  I would say that it is because it is really the only way to do it to get the results that you are looking for with different times.

As to having multiple datasets and having the sources spread - you can think about it the other way.  By having them separated if one of them changes drastically the rest of your report works fine still.  This can actually be what saves you time in the future.  

If he asks to split again, this will not be an entirely new report.  You create the new (third) dataset, and you repoint the queries that used those fields to that new dataset as your data source and the visuals and reports will continue to operate as originally intended.  If certain fields are separated and not a whole dataset (like one of the SQL's) then that will require more work.  

I understand what you are saying about doing something just to do it because you can.  I agree that "easiest" is the best way.  The reason for multiple datasets is based on the needs - if you are refreshing differently you need multiple datasets.  If you aren't you won't need 10 datasets as you theorized.  

In my experience only a few clients have a need for different refresh times for different datasets so I don't think that this is going to be a constant experience so doing it with multiple datasets in this case is a normal solution.

I would appreciate Kudos if my response was helpful. I would also appreciate it if you would Mark this As a Solution if it solved the problem. Thanks!

[danielmillion]

After reading your answer, I do tend to agree with what you say. I really liked your point of view of things.
There is no point to deny clients of something they need and we can provide in a good and professional way.

Thank you @collinq .