fix: Fix ColumnNotFound error selecting len() after semi/anti join

[nameexhaustion]

• Fixes Calling .select(pl.len()) after an anti-join fails to find join columns fails for LazyFrames (but not DataFrames) #21343

Note that there are a few other occurrences of acc_projections.is_empty() - I'm not sure whether those can be problematic:

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 79.94%. Comparing base (4e286d8) to head (9bcbe7d).
Report is 6 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #21355      +/-   ##
==========================================
+ Coverage   79.92%   79.94%   +0.01%     
==========================================
  Files        1596     1596              
  Lines      228745   228623     -122     
  Branches     2616     2618       +2     
==========================================
- Hits       182825   182764      -61     
+ Misses      45321    45260      -61     
  Partials      599      599              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ritchie46]

Ah yes, this is due to a recent optimization that doesn't select (the first column) anymore for select(pl.len()).

[sergiojarala]

Ah yes, this is due to a recent optimization that doesn't select (the first column) anymore for select(pl.len()).

Doesn’t select (the first column) only after anti/semi joins? Or always ? I used to do select(pl.len()) to determine the number of rows in a Lazyframe thinking this would be less expensive than collecting the lazyframe and then do .height()