Sharing Date Table for multiple semantic models in multiple Workspaces (and reports fail)

[Tom_Y]

Hi all,

We have a Date Table saved in Workspace A, and we have Model B in Workspaces B, C in Workspaces C, and then D E F etc.

All the models are now linked to to that same Date Table (as a composite model?)

For consumer to view Report B, C, D, now I need to additionally grant "viewer" to each consumer on Workspace A or there're error in viusals.

The initial plan was that sooner or later we will have our "Organizational Report" so we better have one Date Table for all projects...

What is the industry practice on this kind of Date Table (e.g. One for all or one for each Workspaces?)

(We don't have Premium, we just have Pro)

The question seems silly and I can't find similar topic here or in YouTube, but it's making me worried. It's embarassing when top executive cannot open your BI report...

Many thanks in advance.

[aj1973]

@Tom_Y joining what was said by @nandic a Dataflow for date table entity is a good use case since Dataflows are meant for a unified ETL and reusable purposes. Add a security group to the workspace of the Dataflow as viewer.

[nandic]

@Tom_Y what is source for this Date table?
Best practice would be to create Date table within Dataflow and then just reference it in each dataset. This is the main purpose, when we have tables like Date table which we reference in multiple datasets.
Another approach is just to easily add that date table in each dataset. 
Creating composite models just because of data table is not a good practise.

Cheers,
Nemanja Andic

[Tom_Y]

@nandic @aj1973 Thanks both.

Currently we didn't really touch Dataflows. We only get Pro license (small SME) and some of the semantic models have "computed tables" and we can't refresh it in Dataflows. (we simply make it in BI Desktop, upload to workspaces, and share report to consumers. We didn't even put consumers in Workspaces, but only start to do so when they sometimes fail to view the report)

Can I just put Date Table in Dataflows, and keeping all our composite models in Power BI Desktop + Online Services?

The Date Table is merely a DAX created table and saved separately in a dataset called "Date Table" in one of the Workspaces.

We used to have individual Date Table for each dataset, but when we try to combine datasets into "Organization datasets" we start to struggle on Date Table.

Thanks again.

[aj1973]

No in this case no need for a Dataflow. Since your Date table is a DAX then you can create a pbit power BI template with the dax for date table and provide it to the developers each time they need to build new reports.

[lbendlin]

I have two opposing opinions on that

1.  That's how composite models work. Same with Premium where all users need to "install"  all apps from the participating semantic models. You can choose to "auto install"  apps to avoid that.

2. The Date table is very small and there is no problem to reference it multiple times. You could even import a master date table that controls all the other date tables.  When you have composite models you need to watch out for links with key column cardinality over 50K,

[Tom_Y]

@lbendlin Thanks.

Do you mean/ think sharing an "App" instead of "Report" may solve the problem or do I still need to grant permission to each consumer in each "upper level data set" of those composite models?

The Date Table is merely a DAX created simple table (Date, FY, FQ, sorting order, that's it) and saved separately in a dataset called "Date Table". Currently it seems to be linked to each composite models by "direct query". Do you think I can link them by import?

Thanks again.

[lbendlin]

Yes to both. But if you are a small company then Premium may not be for you.