Reimplementation of NestedFrame.reduce()

[hombit]

Here is a re-implementation of reduce without using df.join and df.assign. It is shorter and runs faster

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 98.76%. Comparing base (aadb12f) to head (fc0b320).
Report is 59 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #96   +/-   ##
=======================================
  Coverage   98.76%   98.76%           
=======================================
  Files          15       15           
  Lines         894      894           
=======================================
  Hits          883      883           
  Misses         11       11           

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

[github-actions]

Before [aadb12f]	After [5fb69b5]	Ratio	Benchmark (Parameter)
31.2±1ms	32.5±2ms	1.04	benchmarks.AssignSingleDfToNestedSeries.time_run
88.2M	90.2M	1.02	benchmarks.NestedFrameQuery.peakmem_run
8.97±0.1ms	9.01±0.1ms	1.01	benchmarks.NestedFrameAddNested.time_run
257M	257M	1	benchmarks.AssignSingleDfToNestedSeries.peakmem_run
84M	84M	1	benchmarks.NestedFrameAddNested.peakmem_run
6.30±0.08ms	6.30±0.1ms	1	benchmarks.NestedFrameQuery.time_run
87.2M	87.2M	1	benchmarks.NestedFrameReduce.peakmem_run
277M	277M	1	benchmarks.ReassignHalfOfNestedSeries.peakmem_run
62.8±2ms	60.0±1ms	0.96	benchmarks.ReassignHalfOfNestedSeries.time_run
5.20±0.03ms	1.06±0ms	0.2	benchmarks.NestedFrameReduce.time_run

Click here to view all benchmarks.

[wilsonbb]

This looks cool, thanks for this!