Remove SortKeyCursor (~5-10% faster merge)

[tustvold]

Which issue does this PR close?

Part of #5882
Relates to #5879

Rationale for this change

Replaces SortKeyCursor with a simplified, crate-private version. I couldn't see an obvious reason why SortKeyCursor was public other than to enable integration tests. Especially now that arrow-row exists, I don't see a compelling reason to keep it around.

merge sorted i64        time:   [8.7181 ms 8.7273 ms 8.7378 ms]
                        change: [-8.6612% -8.5375% -8.3983%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  1 (1.00%) high mild
  2 (2.00%) high severe

sort merge i64          time:   [8.9273 ms 8.9396 ms 8.9521 ms]
                        change: [-8.4047% -8.2575% -8.0995%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

merge sorted f64        time:   [8.6883 ms 8.7002 ms 8.7140 ms]
                        change: [-8.4371% -8.2785% -8.1206%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
  2 (2.00%) high mild
  3 (3.00%) high severe

sort merge f64          time:   [8.9428 ms 8.9550 ms 8.9680 ms]
                        change: [-7.8564% -7.7071% -7.5387%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

merge sorted utf8 low cardinality
                        time:   [6.2272 ms 6.2357 ms 6.2448 ms]
                        change: [-14.595% -14.461% -14.314%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 6 outliers among 100 measurements (6.00%)
  5 (5.00%) high mild
  1 (1.00%) high severe

sort merge utf8 low cardinality
                        time:   [6.8843 ms 6.9109 ms 6.9380 ms]
                        change: [-13.368% -12.957% -12.521%] (p = 0.00 < 0.05)
                        Performance has improved.

merge sorted utf8 high cardinality
                        time:   [10.147 ms 10.224 ms 10.307 ms]
                        change: [-6.2273% -5.3649% -4.5217%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 21 outliers among 100 measurements (21.00%)
  3 (3.00%) high mild
  18 (18.00%) high severe

sort merge utf8 high cardinality
                        time:   [11.191 ms 11.254 ms 11.321 ms]
                        change: [-7.4775% -6.7869% -6.0794%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 2 outliers among 100 measurements (2.00%)
  2 (2.00%) high mild

merge sorted utf8 tuple time:   [17.150 ms 17.169 ms 17.188 ms]
                        change: [-4.9041% -4.7286% -4.5686%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 4 outliers among 100 measurements (4.00%)
  4 (4.00%) high mild

sort merge utf8 tuple   time:   [20.286 ms 20.399 ms 20.511 ms]
                        change: [-5.5443% -4.8463% -4.1378%] (p = 0.00 < 0.05)
                        Performance has improved.

merge sorted utf8 dictionary
                        time:   [5.6368 ms 5.6413 ms 5.6461 ms]
                        change: [-14.343% -14.257% -14.164%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  3 (3.00%) high mild

sort merge utf8 dictionary
                        time:   [5.7874 ms 5.7961 ms 5.8055 ms]
                        change: [-14.409% -14.234% -14.047%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

merge sorted utf8 dictionary tuple
                        time:   [8.3309 ms 8.3364 ms 8.3423 ms]
                        change: [-10.213% -10.122% -10.033%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

sort merge utf8 dictionary tuple
                        time:   [9.6855 ms 9.7650 ms 9.8424 ms]
                        change: [-8.8845% -7.8176% -6.7832%] (p = 0.00 < 0.05)
                        Performance has improved.

merge sorted mixed dictionary tuple
                        time:   [14.514 ms 14.533 ms 14.553 ms]
                        change: [-5.1546% -5.0148% -4.8671%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 3 outliers among 100 measurements (3.00%)
  2 (2.00%) high mild
  1 (1.00%) high severe

sort merge mixed dictionary tuple
                        time:   [15.994 ms 16.073 ms 16.151 ms]
                        change: [-4.7746% -4.1442% -3.4971%] (p = 0.00 < 0.05)
                        Performance has improved.

merge sorted mixed tuple
                        time:   [16.649 ms 16.675 ms 16.702 ms]
                        change: [-4.6957% -4.5328% -4.3548%] (p = 0.00 < 0.05)
                        Performance has improved.
Found 5 outliers among 100 measurements (5.00%)
  4 (4.00%) high mild
  1 (1.00%) high severe

sort merge mixed tuple  time:   [18.427 ms 18.477 ms 18.527 ms]
                        change: [-5.4182% -5.0185% -4.6040%] (p = 0.00 < 0.05)
                        Performance has improved.

What changes are included in this PR?

Are these changes tested?

Are there any user-facing changes?

Yes SortKeyCursor is removed, it is highly unlikely anyone was using this as it is very coupled with SortPreservingMerge

[alamb]

Looks good to me -- thank you @tustvold

cc @yjshen and @richox who may have some history with this code