(fix): setting on Dataset2D "just works" (like Dataframes) and is thoroughly checked

[ilan-gold]

Since we're enforcing two dimensions on the object (in theory), we should probably be super clear about that while also retaining dataframe-like operations i.e., when you try to "set a column" (in this case, xarray data variable that isn't a coord), it "just works" (sets the column and doesn't do anything else). This behavior matches that of pandas in so far as having a named index doesn't mean it can be set/accessed via df["my_named_index"]

I think there are probably other horrifying ways to invalidate this constraint, so I will try to comb through the xarray docs to fix these separately, but trying to keep this scoped to the issue for now

• Closes Fix setting on Dataset2D without specificed coord #1990
• Tests added
• Release note added (or unnecessary)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.11%. Comparing base (ec65081) to head (279e524).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1992      +/-   ##
==========================================
- Coverage   87.11%   85.11%   -2.00%     
==========================================
  Files          46       46              
  Lines        6947     6968      +21     
==========================================
- Hits         6052     5931     -121     
- Misses        895     1037     +142     

Files with missing lines	Coverage Δ
src/anndata/_core/merge.py	85.06% <100.00%> (-8.89%)	⬇️
src/anndata/_core/xarray.py	98.05% <100.00%> (+0.43%)	⬆️

... and 6 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.