I'm talking about the "rich" kind with multiple tables, measures, relationships.
I ask because that's what I'm working on, and it's slow-going.
I started by noticing that the JSON for these looks basically like SSAS tabular bim files. So I first made a tabular model with the tables/relationships/measures/datatypes/stringformatting I wanted, and tried to copy/paste into a C#-driven push dataset.
So now I'm manually picking through the JSON bits, taking out the ones that won't successfully POST (annotations, for example don't seem to translate from tabular to push datasets).
This process seems wildly inefficient. Is there a better way to develop rich data modelling push datasets?