Use take for dictionary like comparisons

[tustvold]

Which issue does this PR close?

Closes #.

Rationale for this change

Other comparison kernels perform an operation on the underlying child array, before taking the result based on the dictionary indexes. Especially for more expensive operations like those involving strings, this can yield significant returns. As an added bonus it results in less codegen

eq dictionary[10] string[4])
                        time:   [358.35 µs 358.43 µs 358.53 µs]
                        change: [-0.9058% -0.7716% -0.5611%] (p = 0.00 < 0.05)
                        Change within noise threshold.
Found 11 outliers among 100 measurements (11.00%)
  4 (4.00%) high mild
  7 (7.00%) high severe

eq_dyn_utf8_scalar dictionary[10] string[4])
                        time:   [53.265 µs 53.284 µs 53.306 µs]
                        change: [-12.029% -11.781% -11.537%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 9 outliers among 100 measurements (9.00%)
  6 (6.00%) high mild
  3 (3.00%) high severe

gt_eq_dyn_utf8_scalar scalar dictionary[10] string[4])
                        time:   [113.34 µs 113.36 µs 113.39 µs]
                        change: [-5.2402% -4.9959% -4.7432%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 16 outliers among 100 measurements (16.00%)
  5 (5.00%) high mild
  11 (11.00%) high severe

like_utf8_scalar_dyn dictionary[10] string[4])
                        time:   [53.307 µs 53.326 µs 53.347 µs]
                        change: [-80.849% -80.811% -80.757%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 10 outliers among 100 measurements (10.00%)
  7 (7.00%) high mild
  3 (3.00%) high severe

ilike_utf8_scalar_dyn dictionary[10] string[4])
                        time:   [53.972 µs 53.990 µs 54.013 µs]
                        change: [-97.799% -97.796% -97.790%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 10 outliers among 100 measurements (10.00%)
  4 (4.00%) high mild
  6 (6.00%) high severe

What changes are included in this PR?

Are there any user-facing changes?

[tustvold]

The duplication is slightly unfortunate, but will hopefully get cleaned up as part of #3296

[tustvold]

This is somewhat unfortunate, but arrow-select is fairly foundational so I'm not too worried

[viirya]

👍