------------------------------------------------------ Good Tuesday morning/afternoon, Given a table for which there is a label that is 1-1 with the PK of its dataset (that is, each value of

cp__{dataset}_id

has a corresponding value for the label), what is the optimal way to name the fields such that the label is displayed as "{dataset} {label name}"? In the docs, they say that given an attribute

{attribute}

, a label should be named

l__{dataset}__{attribute}__{label name}

. Does this apply to PK (

cp__*

fields) as well?

Yes, I think that the same logic which applies to a__ fields should should apply to the cp__* fields as well. I think it is actually

l__{attribute}__{label name}

without the

dataset

part. You can actually from any existing and published model generate the SQL table or view definitions. To do so, go to the Model and in the view mode (not edit mode!) click the three dots in the top right corner, then select “*create output stage*”. In the next step, confirm your data source and whether you want views or tables. It will generate a SQL for you which you can then manually copy, modify and execute in your database. It will generate the exact column names that are required for each dataset. It will NOT execute this SQL automatically against your database so no worries that generating this will break something.

Hey Michal, Thanks for your answer. I'll try your proposed naming convention. FYI in the docs the dataset component is present: