It seems also that there is a performance limit. It does not allow to load the full model (highrise with lots of timber beams and columns) and aborts loading.
One more addition to that …
by interacting with the visuals the model starts to rebuild itself after every big change. It would be interesting if the visual would keep the model in memory. It seems to start loading it from scratch. Especially for medium to big projects, this would be really beneficial. But this is more icing on the cake stuff … you guys are sweeping us off our feet in a regular manner with what is already working
We’re planning on doing some more work on the PowerBI Visual for the next release cycle. This will include the language issues you’re describing above.
If you’re interested in the details, basically we were using the input name naively… but we can switch to use the QueryName, which is language agnostic (or that’s what I’ve gathered…). Hopefully this will allow the visual to work in any language!
As for this
It’s tricky… but we do have some ideas on how to improve things on this side. The main problem is that the viewer uses a local DB by default (which speeds things up considerably) but the PowerBI visual is sandboxed so it cannot use it This is done for very valid security reasons but it still makes me sad.
There are some things we could improve on the viewer side to ease this issue (some kind of new object loader), but this is all very theoretical at this point.
Forgot to answer this. PowerBI has a limitation of loading 30k objects into a visual. There are in theory some pagination strategies but I never got those to work properly.
EDIT: Not sure how massive the buildings you’re loading are… but do note that if you tell the viewer to load an object, it will also load all of it’s children, so you may be able to “reduce” the amount of objects in your table to less than 30K and still see the entire building, but it would really depend on the use-case for your report, and how granular you need the data to be.
This can be done by some filtering after receiving on the DataConnector side. You can also try our (very experimental) structured receive option that may allow for some extra control on receive (the cost of this is you’ll have to build your own data table similar to the GetByUrl result in your query)
This is done for very valid security reasons but it still makes me sad.