Improve performance of trim expressions #2977
Description
What is the problem the feature request solves?
CometStringExpressionBenchmark shows poor performance:
OpenJDK 64-Bit Server VM 11.0.22+7-LTS on Mac OS X 14.6.1
Apple M3 Max
trim: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Spark 10 50 21 0.1 9794.1 1.0X
Comet (Scan) 11 43 31 0.1 10672.9 0.9X
Comet (Scan + Exec) 15 63 23 0.1 14389.7 0.7X
OpenJDK 64-Bit Server VM 11.0.22+7-LTS on Mac OS X 14.6.1
Apple M3 Max
btrim: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Spark 9 33 24 0.1 9219.6 1.0X
Comet (Scan) 27 72 24 0.0 26224.5 0.4X
Comet (Scan + Exec) 15 74 40 0.1 14454.3 0.6X
OpenJDK 64-Bit Server VM 11.0.22+7-LTS on Mac OS X 14.6.1
Apple M3 Max
ltrim: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Spark 10 77 30 0.1 9671.0 1.0X
Comet (Scan) 10 70 19 0.1 9802.0 1.0X
Comet (Scan + Exec) 16 85 26 0.1 15813.0 0.6X
OpenJDK 64-Bit Server VM 11.0.22+7-LTS on Mac OS X 14.6.1
Apple M3 Max
rtrim: Best Time(ms) Avg Time(ms) Stdev(ms) Rate(M/s) Per Row(ns) Relative
------------------------------------------------------------------------------------------------------------------------
Spark 9 26 28 0.1 8771.6 1.0X
Comet (Scan) 9 49 33 0.1 9270.0 0.9X
Comet (Scan + Exec) 14 56 40 0.1 13827.8 0.6X
Describe the potential solution
Review benchmark to make sure it is valid, then try and optimize.
Additional context
No response