feat(planner): implement explain query in verbose mode

[leiysky]

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

Summary

Support explain logical plan with verbose output, which contains relational property information and statistics.

For example:

root@localhost:8000/default> explain(verbose, logical) select * from test;

EXPLAIN(verbose, logical)
SELECT
  *
FROM
  test

-[ EXPLAIN ]-----------------------------------
EvalScalar
├── scalars: [test.a (#0) AS (#0), test.b (#1) AS (#1)]
├── output columns: [default.test.a, default.test.b]
├── outer columns: []
├── used columns: [default.test.a, default.test.b]
├── cardinality: 10.000
├── precise cardinality: 10
├── statistics
│   ├── default.test.a: { min: 0, max: 9, ndv: 10, null count: 0 }
│   └── default.test.b: { min: hello, max: hello, ndv: 10, null count: 0 }
└── Scan
    ├── table: default.test
    ├── filters: []
    ├── order by: []
    ├── limit: NONE
    ├── output columns: [default.test.a, default.test.b]
    ├── outer columns: []
    ├── used columns: [default.test.a, default.test.b]
    ├── cardinality: 10.000
    ├── precise cardinality: 10
    └── statistics
        ├── default.test.a: { min: 0, max: 9, ndv: 10, null count: 0 }
        └── default.test.b: { min: hello, max: hello, ndv: 10, null count: 0 }

23 rows explain in 0.085 sec. Processed 0 rows, 0 B (0 rows/s, 0 B/s)

Also improve the syntax of explain statement by adding explain options, we will prefer to use EXPLAIN(LOGICAL) instead of EXPLAIN RAW.

Currently we supported following explain options:

• VERBOSE: enable verbose output, only works with logical plan
• LOGICAL: print unoptimized logical plan, the same with EXPLAIN RAW
• OPTIMIZED: print optimized logical plan, the same with EXPLAIN OPTIMIZED

In the future we can support FORMAT option to enable different output format of explain statement, e.g. JSON format for strucutral analysis.

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 

[BohuTANG]

Cool.
Btw, we already have explain analyze, syntax like:

explain [analyze] <SQL>

Can we merge this PR feature into the analyze instead of introducing the new explain(verbose, logical)? Or it's hard to choose when to use.

[leiysky]

Cool. Btw, we already have explain analyze, syntax like:

explain [analyze] <SQL>

Can we merge this PR feature into the analyze instead of introducing the new explain(verbose, logical)? Or it's hard to choose when to use.

They are not compatible because explain analyze and explain print PhysicalPlan while explain(logical) prints the logical plan SExpr. They are not even exact one-to-one mapping.

In most of the scenarios explain and explain analyze is enough for users. The explain(verbose, logical) is designed for developers to trace the internal information of optimizer.