Right now in the dataset exploration UI, the ordering of the fields and models are done alphabetically. This is troublesome because it doesn’t allow me to curate/customize what I want to present to the end users.
1/ In a star schema structure, I will want to put the facts model on top (and ideally expanded by default) so that end users’ attention are immediately brought to it.
2/ Within a model, I want the ability to reorder the fields, for various reasons:
Most important fields (maybe measures) can be on top.
I’m used to ID field being at the top of the list
Grouping related fields together
3/ Will be nice to be able to add some Markdown-based README for the dataset, so that analysts can provide instructions to end users on how to use it (key use cases, special logic, etc).
I’d imagine some of these would be quite straightforward? For example, with the models and fields defined in AML, we should just list the fields and models in the order that they are defined in the dataset file and model file?
Agree that more options for presentation would be helpful - for me, by far the most valuable would be the ability to group dimensions or measures together and also to create drill down hierarchies which group into a single dimension group.
I would prefer some form of sort_key or sort_order you apply to the measure/dimension in the model, rather than taking the order in which they appear.
There might be purpose to the order of dimensions, such as existing dimensions, followed by custom dimensions, and this may not be the order I actually want them displayed, but want to keep them in that order in the model.