Optimize PrimitiveBuilder::append_trusted_len_iter

[alamb]

Which issue does this PR close?

• part of [EPIC] Faster performance for parquet predicate evaluation for non selective filters #7456
• Closes Optimize PrimitiveBuilder::append_trusted_len_iter #7591

Rationale for this change

@Dandandan noted on https://github.com/apache/arrow-rs/pull/7513/files#r2119136807:

Looking at the code - this doesn't actually really use the TrustedLen optimization.
I think it could use the same optimization as MutableBuffer::from_trusted_len

What changes are included in this PR?

Use the trusted len iter optimization for extendig PrimitiveBuffers

Are there any user-facing changes?

The only place I can see this used now is take with fixed size arrays. Thus this may make that somewhat faster (I will run some benchmarks)

I think the real value will be in the APIs sketched out in #7513 to append to StringViewBuilder faster

[alamb]

🤖 ./gh_compare_arrow.sh Benchmark Script Running
Linux aal-dev 6.11.0-1013-gcp #13~24.04.1-Ubuntu SMP Wed Apr 2 16:34:16 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
Comparing alamb/from_trusted_len (64c6cdd) to 3681540 diff
BENCH_NAME=take_kernels
BENCH_COMMAND=cargo bench --features=arrow,async,test_common,experimental --bench take_kernels
BENCH_FILTER=
BENCH_BRANCH_NAME=alamb_from_trusted_len
Results will be posted here when complete

[alamb]

🤖: Benchmark completed

Details

group                                                                     alamb_from_trusted_len                 main
-----                                                                     ----------------------                 ----
take bool 1024                                                            1.00   1396.5±2.16ns        ? ?/sec    1.00   1396.5±2.56ns        ? ?/sec
take bool 512                                                             1.00    725.6±0.77ns        ? ?/sec    1.00    726.5±1.06ns        ? ?/sec
take bool null indices 1024                                               1.00      2.5±0.05µs        ? ?/sec    1.08      2.7±0.03µs        ? ?/sec
take bool null values 1024                                                1.06      3.0±0.00µs        ? ?/sec    1.00      2.8±0.00µs        ? ?/sec
take bool null values null indices 1024                                   1.00      5.5±0.07µs        ? ?/sec    1.06      5.9±0.04µs        ? ?/sec
take check bounds i32 1024                                                1.00   1159.0±2.79ns        ? ?/sec    1.01  1172.3±14.19ns        ? ?/sec
take check bounds i32 512                                                 1.06    632.1±1.44ns        ? ?/sec    1.00    597.0±7.56ns        ? ?/sec
take i32 1024                                                             1.00    580.2±1.32ns        ? ?/sec    1.00    579.9±0.74ns        ? ?/sec
take i32 512                                                              1.00    380.3±0.53ns        ? ?/sec    1.00    378.4±0.39ns        ? ?/sec
take i32 null indices 1024                                                1.00    788.6±0.98ns        ? ?/sec    1.01    795.1±2.44ns        ? ?/sec
take i32 null values 1024                                                 1.03      2.1±0.00µs        ? ?/sec    1.00      2.1±0.00µs        ? ?/sec
take i32 null values null indices 1024                                    1.00      3.2±0.04µs        ? ?/sec    1.07      3.4±0.03µs        ? ?/sec
take primitive fsb value len: 12, indices: 1024                           1.00      8.0±0.01µs        ? ?/sec    1.00      8.0±0.05µs        ? ?/sec
take primitive fsb value len: 12, null values, indices: 1024              1.05     10.0±0.07µs        ? ?/sec    1.00      9.5±0.12µs        ? ?/sec
take primitive run logical len: 1024, physical len: 512, indices: 1024    1.02     20.6±0.10µs        ? ?/sec    1.00     20.2±0.08µs        ? ?/sec
take str 1024                                                             1.01     10.8±0.05µs        ? ?/sec    1.00     10.7±0.04µs        ? ?/sec
take str 512                                                              1.03      5.5±0.02µs        ? ?/sec    1.00      5.3±0.01µs        ? ?/sec
take str null indices 1024                                                1.00      9.2±0.03µs        ? ?/sec    1.23     11.2±0.04µs        ? ?/sec
take str null indices 512                                                 1.00      3.5±0.02µs        ? ?/sec    1.43      5.0±0.03µs        ? ?/sec
take str null values 1024                                                 1.00     10.1±0.04µs        ? ?/sec    1.11     11.2±0.06µs        ? ?/sec
take str null values null indices 1024                                    1.00      9.1±0.06µs        ? ?/sec    1.04      9.5±0.05µs        ? ?/sec
take stringview 1024                                                      1.12    854.0±0.92ns        ? ?/sec    1.00    765.4±2.47ns        ? ?/sec
take stringview 512                                                       1.00    473.2±1.04ns        ? ?/sec    1.02    483.8±3.55ns        ? ?/sec
take stringview null indices 1024                                         1.02  1390.9±23.41ns        ? ?/sec    1.00   1359.2±7.85ns        ? ?/sec
take stringview null indices 512                                          1.04    689.5±1.62ns        ? ?/sec    1.00    662.0±1.13ns        ? ?/sec
take stringview null values 1024                                          1.01      2.1±0.00µs        ? ?/sec    1.00      2.0±0.00µs        ? ?/sec
take stringview null values null indices 1024                             1.00      3.7±0.06µs        ? ?/sec    1.05      3.8±0.07µs        ? ?/sec

[alamb]

I need to spend some time making sure it is actually faster

[alamb]

I think we are going to go with a different approach (use a Vec instead)