refactor(query): rewrite function call expr to cast expr

[forsaken628]

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

Summary

A series of refactorings/optimizations have been done around ExprVisitor.

1. functions like to_string are rewritten as cast expressions, which makes it easier to add other expression logic later on
2. check_function will prioritize cast wrap on const parameters.
3. Expr display provides the function to display ids.
4. add ut test_type_check
5. reduce the memory size of enum Expr from 288 bytes -> 128 bytes.

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 

[sundy-li]

So to_xxx will be checked as CastExpr in type checker and checked back into to_xxx function in runtime.

LGTM.

Do you perf the following SQLs with the previous version?

select  to_string(number), to_int64(number) from numbers(1000000000) ignore_result;

select  cast(number as string), cast(number as int64) from numbers(1000000000) ignore_result;

[forsaken628]

v1.2.716-nightly-979dab1299
--
v1.2.718-nightly-208d6feede

explain analyze select  to_string(number), to_int64(number) from numbers(1000000000) ignore_result;

EvalScalar
├── output columns: [to_string(number) (#1), to_int64(number) (#2)]
├── expressions: [to_string(numbers.number (#0)), to_int64(numbers.number (#0))]
├── estimated rows: 1000000000.00
├── cpu time: 14.703560079s
├── output rows: 1 billion
├── output bytes: 26.90 GiB
--
EvalScalar
├── output columns: [to_string(number) (#1), to_int64(number) (#2)]
├── expressions: [CAST(numbers.number (#0) AS String), CAST(numbers.number (#0) AS Int64)]
├── estimated rows: 1000000000.00
├── cpu time: 12.791843634s
├── output rows: 1 billion
├── output bytes: 26.90 GiB

explain analyze select cast(number as string), cast(number as int64) from numbers(1000000000) ignore_result;

EvalScalar
├── output columns: [CAST(number AS STRING) (#1), CAST(number AS Int64) (#2)]
├── expressions: [to_string(numbers.number (#0)), to_int64(numbers.number (#0))]
├── estimated rows: 1000000000.00
├── cpu time: 14.755832024s
├── output rows: 1 billion
├── output bytes: 26.90 GiB
--
EvalScalar
├── output columns: [CAST(number AS STRING) (#1), CAST(number AS Int64) (#2)]
├── expressions: [CAST(numbers.number (#0) AS String), CAST(numbers.number (#0) AS Int64)]
├── estimated rows: 1000000000.00
├── cpu time: 12.869861283s
├── output rows: 1 billion
├── output bytes: 26.90 GiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select to_string(l_orderkey) from t;

EvalScalar
├── output columns: [to_string(l_orderkey) (#18)]
├── expressions: [to_string(t.l_orderkey (#0))]
├── estimated rows: 18003645.00
├── cpu time: 344.298703ms
├── output rows: 30.01 million
├── output bytes: 538.40 MiB
--
EvalScalar
├── output columns: [to_string(l_orderkey) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS String)]
├── estimated rows: 18003645.00
├── cpu time: 341.178704ms
├── output rows: 30.01 million
├── output bytes: 538.40 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select l_orderkey::string from t;

EvalScalar
├── output columns: [l_orderkey::STRING (#18)]
├── expressions: [to_string(t.l_orderkey (#0))]
├── estimated rows: 18003645.00
├── cpu time: 345.455491ms
├── output rows: 30.01 million
├── output bytes: 538.40 MiB
--
EvalScalar
├── output columns: [l_orderkey::STRING (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS String)]
├── estimated rows: 18003645.00
├── cpu time: 343.893416ms
├── output rows: 30.01 million
├── output bytes: 538.40 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select cast(l_orderkey as string) from t;

EvalScalar
├── output columns: [CAST(l_orderkey AS STRING) (#18)]
├── expressions: [to_string(t.l_orderkey (#0))]
├── estimated rows: 18003645.00
├── cpu time: 353.688089ms
├── output rows: 30.01 million
├── output bytes: 538.40 MiB
--
EvalScalar
├── output columns: [CAST(l_orderkey AS STRING) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS String)]
├── estimated rows: 18003645.00
├── cpu time: 346.544552ms
├── output rows: 30.01 million
├── output bytes: 538.40 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select cast(l_orderkey as int64 null) from t;

EvalScalar
├── output columns: [CAST(l_orderkey AS Int64 NULL) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS Int64 NULL)]
├── estimated rows: 18003645.00
├── cpu time: 7.222148ms
├── output rows: 30.01 million
├── output bytes: 232.51 MiB
--
EvalScalar
├── output columns: [CAST(l_orderkey AS Int64 NULL) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS Int64 NULL)]
├── estimated rows: 18003645.00
├── cpu time: 3.83956ms
├── output rows: 30.01 million
├── output bytes: 232.51 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select cast(l_orderkey as uint64 null) from t;

EvalScalar
├── output columns: [CAST(l_orderkey AS UInt64 NULL) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS UInt64 NULL)]
├── estimated rows: 18003645.00
├── cpu time: 153.676638ms
├── output rows: 30.01 million
├── output bytes: 232.51 MiB
--
EvalScalar
├── output columns: [CAST(l_orderkey AS UInt64 NULL) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS UInt64 NULL)]
├── estimated rows: 18003645.00
├── cpu time: 79.140626ms
├── output rows: 30.01 million
├── output bytes: 232.51 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select to_uint64(l_orderkey) from t;

EvalScalar
├── output columns: [to_uint64(l_orderkey) (#18)]
├── expressions: [to_uint64(t.l_orderkey (#0))]
├── estimated rows: 18003645.00
├── cpu time: 70.598677ms
├── output rows: 30.01 million
├── output bytes: 228.93 MiB
--
EvalScalar
├── output columns: [to_uint64(l_orderkey) (#18)]
├── expressions: [CAST(t.l_orderkey (#0) AS UInt64)]
├── estimated rows: 18003645.00
├── cpu time: 76.400565ms
├── output rows: 30.01 million
├── output bytes: 228.93 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select to_int64(to_uint32(l_orderkey)) from t;

EvalScalar
├── output columns: [to_int64(to_uint32(l_orderkey)) (#18)]
├── expressions: [to_int64(to_uint32(t.l_orderkey (#0)))]
├── estimated rows: 18003645.00
├── cpu time: 84.32972ms
├── output rows: 30.01 million
├── output bytes: 228.93 MiB
--
EvalScalar
├── output columns: [to_int64(to_uint32(l_orderkey)) (#18)]
├── expressions: [CAST(CAST(t.l_orderkey (#0) AS UInt32) AS Int64)]
├── estimated rows: 18003645.00
├── cpu time: 99.113361ms
├── output rows: 30.01 million
├── output bytes: 228.93 MiB

explain analyze with t as (select l_orderkey,unnest([1,2,3,4,5]) from lineitem) select to_int64(l_orderkey::uint32 null) from t;

EvalScalar
├── output columns: [to_int64(l_orderkey::UInt32 NULL) (#18)]
├── expressions: [to_int64(CAST(t.l_orderkey (#0) AS UInt32 NULL))]
├── estimated rows: 18003645.00
├── cpu time: 163.237386ms
├── output rows: 30.01 million
├── output bytes: 232.51 MiB
--
EvalScalar
├── output columns: [to_int64(l_orderkey::UInt32 NULL) (#18)]
├── expressions: [CAST(CAST(t.l_orderkey (#0) AS UInt32 NULL) AS Int64 NULL)]
├── estimated rows: 18003645.00
├── cpu time: 91.626866ms
├── output rows: 30.01 million
├── output bytes: 232.51 MiB