Now that there's a "Features" section in the docs, it might be helpful to cover the dy.Collection class in there, with a focus more on "what are some common workflows" (as opposed to the existing example in "Real-world Example", and the noisy API documentation generated for that class).
Some question to think about that might be relevant:
- How do you define foreign key relationships in a
dy.Collection data model? Is it done automatically by column name? Is there a way to do it manually?
- How do you specify the cardinality of join validations (e.g., 1:1, 1:m, etc.).
- How do you specify whether nulls are treated as a "value" in the join, vs. the absense of data? SQL always takes the former approach, but Polars is nice in that it lets you select whether null is a value itself or a marker for missing/unknown data (e.g., by its
nulls_equal= argument).
Now that there's a "Features" section in the docs, it might be helpful to cover the
dy.Collectionclass in there, with a focus more on "what are some common workflows" (as opposed to the existing example in "Real-world Example", and the noisy API documentation generated for that class).Some question to think about that might be relevant:
dy.Collectiondata model? Is it done automatically by column name? Is there a way to do it manually?nulls_equal=argument).