perf: Batch parquet integer decoding

[coastalwhite]

No description provided.

[ritchie46]

Any benchmark numbers? ^^

[coastalwhite]

For single-threaded cold runs:

Benchmark 1: ./plparbench-before
  Time (mean ± σ):     27.046 s ±  0.188 s    [User: 12.426 s, System: 10.449 s]
  Range (min … max):   26.731 s … 27.222 s    5 runs

Benchmark 2: ./plparbench-after
  Time (mean ± σ):     24.166 s ±  0.195 s    [User: 10.760 s, System: 10.243 s]
  Range (min … max):   23.976 s … 24.488 s    5 runs

Summary
  ./plparbench-after ran
    1.12 ± 0.01 times faster than ./plparbench-before

For multi-threaded cold runs the performance gain if any is negledable, since there are other columns that take longer to decode.

[codecov]

Codecov Report

Attention: Patch coverage is 79.31034% with 6 lines in your changes missing coverage. Please review.

Project coverage is 80.44%. Comparing base (7015663) to head (66e1fb7).

Files	Patch %	Lines
...quet/src/arrow/read/deserialize/primitive/basic.rs	33.33%	6 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #17734   +/-   ##
=======================================
  Coverage   80.43%   80.44%           
=======================================
  Files        1502     1502           
  Lines      197035   197041    +6     
  Branches     2794     2794           
=======================================
+ Hits       158482   158502   +20     
+ Misses      38039    38025   -14     
  Partials      514      514           

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