feat: estimate selectivity by table sample

[xudong963]

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

Summary

Support estimating filter selectivity by table sample(0.2%).

The dynamic sample may be increased runtime for the query, so to avoid performance downgrading, add a time budget to limit the time, see the setting dynamic_sample_time_budget_ms, default is 0 ms.

For example, SQL from join order benchmark

Use sample:

EXPLAIN ANALYZE PARTIAL
SELECT
  *
FROM
  movie_companies AS mc,
  movie_info_idx AS mi_idx
WHERE
  mc.movie_id = mi_idx.movie_id
  AND mc.note NOT LIKE '%(as Metro-Goldwyn-Mayer Pictures)%'
  AND (
    mc.note LIKE '%(co-production)%'
    OR mc.note LIKE '%(presents)%'
  )

-[ EXPLAIN ]-----------------------------------
HashJoin
├── estimated rows: 84512.79
├── output rows: 62.66 thousand
├── Filter
│   ├── filters: [is_true(NOT like(mc.note (#4), '%(as Metro-Goldwyn-Mayer Pictures)%')), is_true(like(mc.note (#4), '%(co-production)%') OR like(mc.note (#4), '%(presents)%'))]
│   ├── estimated rows: 28495.95
│   ├── output rows: 28.89 thousand
│   └── TableScan
│       ├── table: default.imdb.movie_companies
│       ├── estimated rows: 2609129.00
│       └── output rows: 2.61 million
└── TableScan
    ├── table: default.imdb.movie_info_idx
    ├── estimated rows: 1380035.00
    └── output rows: 1.38 million

15 rows explain in 1.831 sec. Processed 0 rows, 0 B (0 rows/s, 0 B/s)

For filter, estimated rows is similar to output rows(real rows).

No sample

-[ EXPLAIN ]-----------------------------------
HashJoin
├── estimated rows: 15584.87
├── output rows: 62.66 thousand
├── Filter
│   ├── filters: [is_true(NOT like(mc.note (#4), '%(as Metro-Goldwyn-Mayer Pictures)%')), is_true(like(mc.note (#4), '%(co-production)%') OR like(mc.note (#4), '%(presents)%'))]
│   ├── estimated rows: 5254.89
│   ├── output rows: 28.89 thousand
│   └── TableScan
│       ├── table: default.imdb.movie_companies
│       ├── estimated rows: 2609129.00
│       └── output rows: 2.61 million
└── TableScan
    ├── table: default.imdb.movie_info_idx
    ├── estimated rows: 1380035.00
    └── output rows: 1.38 million

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