feat(types): introduce Timestamptz type

[wangrunji0408]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

This PR introduces Timestamptz type and array. It's another attempt after #7084.

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
• My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
• All checks passed in ./risedev check (or alias, ./risedev c)
• My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

Documentation

• My PR contains user-facing changes.

[BugenZhao]

May I ask whether we're going to unify logical types (DataType) and physical types (ScalarImpl::..) after this PR?

---

Seems still impossible since there could some metadata in DataType like the field name of structs. But we can now establish 1v1 relationship from logical types to physical types.

[wangrunji0408]

May I ask whether we're going to unify logical types (DataType) and physical types (ScalarImpl::..) after this PR?

Seems still impossible since there could some metadata in DataType like the field name of structs. But we can now establish 1v1 relationship from logical types to physical types.

Exactly. Even though we have a 1-1 mapping between the variants of DataType and ScalarImpl (which is good), for some types it still depends on the metadata in DataType to interpret values in ScalarImpl.

[xiangjinwu]

One example is ScalarImpl::List with a ListValue containing [] or [None]. It is still necessary to have DataType to know whether it is int[] or varchar[] or int[][].

[xiangjinwu]

Oh sorry the build failure is due to an uncaught conflict with #10462. Should be an easy fix. I am reviewing the other parts

[xiangjinwu]

Rest lgtm. Thank you!

[codecov]

Codecov Report

Merging #10477 (fe3ec24) into main (296348a) will decrease coverage by 0.01%.
The diff coverage is 65.37%.

@@            Coverage Diff             @@
##             main   #10477      +/-   ##
==========================================
- Coverage   70.16%   70.15%   -0.01%     
==========================================
  Files        1295     1296       +1     
  Lines      221572   221690     +118     
==========================================
+ Hits       155455   155527      +72     
- Misses      66117    66163      +46     

Flag	Coverage Δ
rust	70.15% <65.37%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/common/src/array/arrow.rs	67.00% <0.00%> (-1.72%)	⬇️
src/common/src/array/chrono_array.rs	100.00% <ø> (ø)
src/common/src/array/proto_reader.rs	87.17% <0.00%> (-1.71%)	⬇️
src/common/src/array/stream_chunk.rs	80.96% <ø> (ø)
src/common/src/hash/key.rs	80.67% <0.00%> (-1.19%)	⬇️
src/common/src/test_utils/rand_chunk.rs	79.16% <0.00%> (ø)
src/common/src/types/datetime.rs	62.74% <ø> (+0.72%)	⬆️
src/common/src/types/to_text.rs	74.52% <ø> (-2.34%)	⬇️
src/connector/src/parser/csv_parser.rs	87.09% <0.00%> (ø)
src/expr/macro/src/lib.rs	94.73% <ø> (ø)
... and 30 more

... and 2 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more