macrostrat.unit_boundaries missing explicit foreign keys

[cambro]

In the current dump, unit_boundaries.unit_id is mapped to units.id but unit_id_2 is not. It should be mapped to units.id as well. It would be desirable to name these fields in an informative way too. I regret my choices here! (unit_above and unit_below would be way more intuitive).

t1 should also be mapped to intervals.id.
section_id should also be mapped to sections.id

[davenquinn]

Yes, these are all omissions that can be corrected.

I agree with the field name change; I will take a look at the API to see if we can implement it easily without breaking changes on the user side.

One design idea:

• Rename the unit_boundaries table to unit_surfaces and make the name changes
• Create a view unit_boundaries that mirrors the unit_surfaces table, mapping new field names to the old ones and allowing the API/current scripts to function as they currently do

I like the unit_surfaces name because it is parallel with a new table I am envisioning called surfaces or time_surfaces, which would hold surfaces that correlate across multiple columns/unit pairs.