perf: improve performance of lpad/rpad by reusing buffers

[viirya]

Optimized lpad and rpad functions to eliminate per-row allocations by reusing buffers for graphemes and fill characters.

The previous implementation allocated new Vec<&str> for graphemes and Vec for fill characters on every row, which was inefficient. This optimization introduces reusable buffers that are allocated once and cleared/refilled for each row.

Changes:

• lpad: Added graphemes_buf and fill_chars_buf outside loops, clear and refill per row instead of allocating new Vec each time
• rpad: Added graphemes_buf outside loops to reuse across iterations
• Both functions now allocate buffers once and reuse them for all rows
• Buffers are cleared and reused for each row via .clear() and .extend()

Optimization impact:

• For lpad with fill parameter: Eliminates 2 Vec allocations per row (graphemes + fill_chars)
• For lpad without fill: Eliminates 1 Vec allocation per row (graphemes)
• For rpad: Eliminates 1 Vec allocation per row (graphemes)

This optimization is particularly effective for:

• Large arrays with many rows
• Strings with multiple graphemes (unicode characters)
• Workloads with custom fill patterns

Benchmark results comparing main vs optimized branch:

lpad benchmarks:

• size=1024, str_len=5, target=20: 116.53 µs -> 63.226 µs (45.7% faster)
• size=1024, str_len=20, target=50: 314.07 µs -> 190.30 µs (39.4% faster)
• size=4096, str_len=5, target=20: 467.35 µs -> 261.29 µs (44.1% faster)
• size=4096, str_len=20, target=50: 1.2286 ms -> 754.24 µs (38.6% faster)

rpad benchmarks:

• size=1024, str_len=5, target=20: 113.89 µs -> 72.645 µs (36.2% faster)
• size=1024, str_len=20, target=50: 313.68 µs -> 202.98 µs (35.3% faster)
• size=4096, str_len=5, target=20: 456.08 µs -> 295.57 µs (35.2% faster)
• size=4096, str_len=20, target=50: 1.2523 ms -> 818.47 µs (34.6% faster)

Overall improvements: 35-46% faster across all workloads

Which issue does this PR close?

• Closes #.

Rationale for this change

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

[Copilot]

Pull request overview

This PR optimizes the performance of the lpad and rpad functions by eliminating per-row Vec allocations through buffer reuse. The optimization introduces reusable buffers for graphemes and fill characters that are allocated once outside the loops and cleared/refilled for each row, rather than allocating new vectors on every iteration.

Key changes:

• Added buffer reuse optimization for graphemes in both lpad and rpad functions
• Added buffer reuse optimization for fill characters in lpad function
• Enhanced benchmarks with more comprehensive test cases covering different array sizes, string lengths, and data types

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
datafusion/functions/src/unicode/lpad.rs	Implements buffer reuse for graphemes and fill characters by declaring Vec buffers outside loops and using clear()/extend() pattern instead of per-row allocations
datafusion/functions/src/unicode/rpad.rs	Implements buffer reuse for graphemes using the same clear()/extend() pattern; however, missing the fill_chars_buf optimization that was implemented in lpad
datafusion/functions/benches/pad.rs	Restructures benchmarks with more comprehensive test coverage including different array sizes (1024, 4096), string lengths (5, 20), target lengths (20, 50), and data types (Utf8, StringView)

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The fill characters should be pre-collected into a reusable buffer similar to how it's done in lpad.rs. Currently, fill.chars() is being called on every row which allocates a new iterator each time. This misses an optimization opportunity mentioned in the PR description: "For lpad with fill parameter: Eliminates 2 Vec allocations per row (graphemes + fill_chars)". The rpad function should also benefit from the same fill_chars_buf optimization by declaring a fill_chars_buf at line 219 and reusing it by clearing and extending with fill.chars() before this usage.