IN LIST: add string-view filters for Utf8View and BinaryView

[geoffreyclaude]

Which issue does this PR close?

• Part of Further improve performance of IN list evaluation #19241.
• Stacked on IN LIST: add direct-probe hash filter for large primitive lists #23015.
• Next in stack: IN LIST: reinterpret FixedSizeBinary for primitive fast paths #23018.
• Extracted from Optimize IN performance with specialized implementations #19390.

Rationale for this change

String and binary values are different from integers because comparing the full bytes can be more expensive. A string may be long, and many rows may not match at all.

Arrow's Utf8View and BinaryView layouts store useful summary information directly in each view value: the length and a prefix of the bytes. That gives us a cheap first question:

Could this value possibly match anything in the IN list?

If the length and prefix do not match, the answer is definitely no, and we avoid comparing the full value. If they do match, the value is only a candidate. Long values are then verified with an exact byte comparison before returning true.

For short Utf8View strings, the whole value fits inline in the view itself, so the primitive fast paths can be reused directly.

What changes are included in this PR?

• Adds short-string Utf8View branchless/hash filters by viewing inline string views as 16-byte values.
• Adds ByteViewMaskedFilter for mixed-length Utf8View and BinaryView arrays.
• Uses length/prefix-style masked view values for fast rejection.
• Performs exact full-byte verification for long candidates before reporting a match.
• Adds a result-building helper that skips expensive membership checks for null needles.
• Adds focused coverage for short sliced views and long-string prefix collisions.

Are these changes tested?

Yes.

• cargo fmt --all --check
• cargo test -p datafusion-physical-expr reinterpreted_ --lib
• cargo test -p datafusion-physical-expr utf8view_hash_filter_handles_short_slices --lib
• cargo test -p datafusion-physical-expr byte_view_masked_filter_verifies_long_string_matches --lib
• cargo test -p datafusion-physical-expr in_list_string_types --lib
• cargo test -p datafusion-physical-expr in_list_binary_types --lib
• cargo clippy -p datafusion-physical-expr --all-targets --all-features -- -D warnings

Are there any user-facing changes?

No. This is an internal performance optimization only.

Local benchmark snapshot

Benchmark command:

cargo bench -p datafusion-physical-expr --profile release-nonlto --bench in_list_strategy -- --save-baseline <name>

Method: compare adjacent saved baselines using raw Criterion sample minima (min(time / iters)). Lower is better; changes within +/-5% are treated as noise.

Compared baselines: #23015 -> #23016

Relevant scope: Utf8View and nullable Utf8View rows.

Summary: 34 relevant rows, 32 faster, 2 slower, 0 within +/-5%.

Largest relevant deltas:

Benchmark	Before	After	Change
nulls/utf8view/short_8b/list=16/match=50%/nulls=20%	61.49 us	12.51 us	-79.6% (4.91x faster)
nulls/utf8view/short_8b/list=16/match=50%/nulls=50%	60.93 us	13.00 us	-78.7% (4.69x faster)
nulls/utf8view/short_8b/list=16/match=50%/nulls=20%/NOT_IN	58.66 us	13.08 us	-77.7% (4.49x faster)
utf8view/short_8b/list=256/match=50%	48.84 us	11.56 us	-76.3% (4.23x faster)
utf8view/short_8b/list=16/match=50%	48.24 us	11.49 us	-76.2% (4.20x faster)
utf8view/len_12b/list=16/match=50%	48.64 us	11.74 us	-75.9% (4.14x faster)
utf8view/len_12b/list=64/match=50%	47.46 us	11.53 us	-75.7% (4.12x faster)
utf8view/short_8b/list=64/match=50%	47.34 us	12.20 us	-74.2% (3.88x faster)
utf8view/short_8b/list=4/match=50%	45.63 us	12.19 us	-73.3% (3.74x faster)
utf8view/mixed_len/list=16/match=50%	98.48 us	27.32 us	-72.3% (3.60x faster)
utf8view/shared_prefix/pfx=12/list=32/match=0%	38.92 us	11.61 us	-70.2% (3.35x faster)
utf8view/shared_prefix/pfx=16/list=64/match=0%	38.46 us	11.68 us	-69.6% (3.29x faster)
utf8view/mixed_len/list=64/match=0%	41.15 us	13.53 us	-67.1% (3.04x faster)
utf8view/mixed_len/list=16/match=0%	40.41 us	13.59 us	-66.4% (2.97x faster)
utf8view/long_24b/list=4/match=0%	37.46 us	13.03 us	-65.2% (2.87x faster)

Full relevant table (34 rows)

Benchmark	Before	After	Change
nulls/utf8view/long_24b/list=16/match=50%/nulls=20%	87.80 us	64.71 us	-26.3% (1.36x faster)
nulls/utf8view/short_8b/list=16/match=50%/nulls=20%	61.49 us	12.51 us	-79.6% (4.91x faster)
nulls/utf8view/short_8b/list=16/match=50%/nulls=20%/NOT_IN	58.66 us	13.08 us	-77.7% (4.49x faster)
nulls/utf8view/short_8b/list=16/match=50%/nulls=50%	60.93 us	13.00 us	-78.7% (4.69x faster)
utf8view/len_12b/list=16/match=0%	20.27 us	10.97 us	-45.9% (1.85x faster)
utf8view/len_12b/list=16/match=50%	48.64 us	11.74 us	-75.9% (4.14x faster)
utf8view/len_12b/list=64/match=0%	20.43 us	10.72 us	-47.5% (1.91x faster)
utf8view/len_12b/list=64/match=50%	47.46 us	11.53 us	-75.7% (4.12x faster)
utf8view/long_24b/list=16/match=0%	37.39 us	14.08 us	-62.3% (2.65x faster)
utf8view/long_24b/list=16/match=50%	87.66 us	74.12 us	-15.5% (1.18x faster)
utf8view/long_24b/list=256/match=0%	38.06 us	18.09 us	-52.5% (2.10x faster)
utf8view/long_24b/list=256/match=50%	88.59 us	97.05 us	+9.5% (1.10x slower)
utf8view/long_24b/list=4/match=0%	37.46 us	13.03 us	-65.2% (2.87x faster)
utf8view/long_24b/list=4/match=50%	86.98 us	74.40 us	-14.5% (1.17x faster)
utf8view/long_24b/list=64/match=0%	38.03 us	19.04 us	-49.9% (2.00x faster)
utf8view/long_24b/list=64/match=50%	87.42 us	97.52 us	+11.6% (1.12x slower)
utf8view/mixed_len/list=16/match=0%	40.41 us	13.59 us	-66.4% (2.97x faster)
utf8view/mixed_len/list=16/match=50%	98.48 us	27.32 us	-72.3% (3.60x faster)
utf8view/mixed_len/list=64/match=0%	41.15 us	13.53 us	-67.1% (3.04x faster)
utf8view/mixed_len/list=64/match=50%	111.64 us	43.85 us	-60.7% (2.55x faster)
utf8view/shared_prefix/pfx=12/list=32/match=0%	38.92 us	11.61 us	-70.2% (3.35x faster)
utf8view/shared_prefix/pfx=12/list=32/match=50%	91.20 us	74.16 us	-18.7% (1.23x faster)
utf8view/shared_prefix/pfx=16/list=64/match=0%	38.46 us	11.68 us	-69.6% (3.29x faster)
utf8view/shared_prefix/pfx=16/list=64/match=50%	88.41 us	76.78 us	-13.2% (1.15x faster)
utf8view/shared_prefix/pfx=8/list=16/match=0%	30.97 us	11.68 us	-62.3% (2.65x faster)
utf8view/shared_prefix/pfx=8/list=16/match=50%	79.36 us	63.75 us	-19.7% (1.24x faster)
utf8view/short_8b/list=16/match=0%	20.58 us	10.78 us	-47.6% (1.91x faster)
utf8view/short_8b/list=16/match=50%	48.24 us	11.49 us	-76.2% (4.20x faster)
utf8view/short_8b/list=256/match=0%	20.90 us	10.82 us	-48.2% (1.93x faster)
utf8view/short_8b/list=256/match=50%	48.84 us	11.56 us	-76.3% (4.23x faster)
utf8view/short_8b/list=4/match=0%	20.30 us	12.15 us	-40.1% (1.67x faster)
utf8view/short_8b/list=4/match=50%	45.63 us	12.19 us	-73.3% (3.74x faster)
utf8view/short_8b/list=64/match=0%	20.82 us	11.11 us	-46.6% (1.87x faster)
utf8view/short_8b/list=64/match=50%	47.34 us	12.20 us	-74.2% (3.88x faster)