refactor(query): sort kernel optimization

[sundy-li]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

1. Introduce ColumnVisitor trait and SortCompare
2. Optimize kernel sort performance (Now we are using column vertical sort, which removes virtual call in sort compare functions)

Sort Performance with 2~3 order column test example, gain about 40%~50%

select l_partkey, l_returnflag from (select * from tpch_test.lineitem limit 10000000) order by l_partkey desc, l_returnflag asc ignore_result;

Tests

• Unit Test
• Logic Test
• Benchmark Test
• No Test - Explain why

Type of change

• Bug Fix (non-breaking change which fixes an issue)
• New Feature (non-breaking change which adds functionality)
• Breaking Change (fix or feature that could cause existing functionality not to work as expected)
• Documentation Update
• Refactoring
• Performance Improvement
• Other (please describe):

---

This change is 

[forsaken628]

Reviewed 24 of 26 files at r1, 2 of 2 files at r2, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @sundy-li)