[andygrove]

Which issue does this PR close?

Closes #.

Rationale for this change

The row_step optimization uses an inner loop processing 10 rows at a time and downcast the builders every 10 rows. It does not seem to do anything to actually improve performance, so I removed it. The code is simpler to understand now and should be faster.

Before:

row_array_conversion/row_to_array
                        time:   [3.9986 ms 4.0807 ms 4.1705 ms]
Found 22 outliers among 100 measurements (22.00%)
  10 (10.00%) high mild
  12 (12.00%) high severe

After:

row_array_conversion/row_to_array
                        time:   [3.6814 ms 3.8756 ms 4.0666 ms]
                        change: [−9.9184% −5.0275% +0.4773%] (p = 0.06 > 0.05)
                        No change in performance detected.

What changes are included in this PR?

How are these changes tested?

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.41%. Comparing base (f09f8af) to head (d83a72d).
⚠️ Report is 771 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #2920      +/-   ##
============================================
+ Coverage     56.12%   59.41%   +3.28%     
- Complexity      976     1379     +403     
============================================
  Files           119      167      +48     
  Lines         11743    15353    +3610     
  Branches       2251     2548     +297     
============================================
+ Hits           6591     9122    +2531     
- Misses         4012     4945     +933     
- Partials       1140     1286     +146     

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

🚀 New features to boost your workflow:

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

[comphead]

Thanks @andygrove that looks like a core change, and looks more easy to read and simpler. Interesting there is no performance gain on dropping the loop