Making sd_format show geospatial values in arrays and structs properly

[Kontinuation]

This patch revamped the implementation of sd_format to work on any input types, and format the spatial columns recursively while leaving non-spatial columns untouched.

Before applying this patch:

> SELECT [ST_POINT(1,2), ST_POINT(3,4)];
┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                                                                         make_array(st_point(Int64(1),Int64(2)),st_point(Int64(3),Int64(4)))                                                                        │
│ list(field { name: "item", data_type: struct([field { name: "geoarrow.wkb", data_type: binary, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {"arrow:extension:name": "geoarrow.wkb", "arrow:exte… │
╞════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
│ [{geoarrow.wkb: 0101000000000000000000f03f0000000000000040}, {geoarrow.wkb: 010100000000000000000008400000000000001040}]                                                                                           │
└────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘

> SELECT struct(ST_POINT(3,4) as a , ST_POINT(3,4) as b)
┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                    named_struct(Utf8("a"),st_point(Int64(3),Int64(4)),Utf8("b"),st_point(Int64(3),Int64(4)))                   │
│                              struct(a struct(geoarrow.wkb binary), b struct(geoarrow.wkb binary))                              │
╞════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
│ {a: {geoarrow.wkb: 010100000000000000000008400000000000001040}, b: {geoarrow.wkb: 010100000000000000000008400000000000001040}} │
└────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘

After applying this patch:

> SELECT [ST_POINT(1,2), ST_POINT(3,4)];
┌───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                                                                        make_array(st_point(Int64(1),Int64(2)),st_point(Int64(3),Int64(4)))                                                                        │
│ list(field { name: "item", data_type: struct([field { name: "geoarrow.wkb", data_type: binary, nullable: true, dict_id: 0, dict_is_ordered: false, metadata: {"arrow:extension:name": "geoarrow.wkb", "arrow:ext… │
╞═══════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════════╡
│ [POINT(1 2), POINT(3 4)]                                                                                                                                                                                          │
└───────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┘

> SELECT struct(ST_POINT(3,4) as a , ST_POINT(3,4) as b);
┌───────────────────────────────────────────────────────────────────────────────────────────┐
│ named_struct(Utf8("a"),st_point(Int64(3),Int64(4)),Utf8("b"),st_point(Int64(3),Int64(4))) │
│            struct(a struct(geoarrow.wkb binary), b struct(geoarrow.wkb binary))           │
╞═══════════════════════════════════════════════════════════════════════════════════════════╡
│ {a: POINT(3 4), b: POINT(3 4)}                                                            │
└───────────────────────────────────────────────────────────────────────────────────────────┘

[Copilot]

Pull Request Overview

This PR revamps the sd_format function to properly handle geospatial values within complex nested data structures (arrays and structs) by recursively formatting spatial columns while leaving non-spatial columns unchanged.

• Refactored implementation to use a single kernel that handles all data types rather than separate kernels for default and geometry types
• Added recursive formatting support for nested structures (structs, lists, and list views)
• Enhanced type system to properly transform geospatial types to UTF8 while preserving other data types

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 5 comments.

File	Description
rust/sedona-functions/src/sd_format.rs	Complete rewrite of sd_format implementation with recursive type handling and comprehensive test coverage
rust/sedona-expr/src/scalar_udf.rs	Added is_any() matcher to support accepting any argument type

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[paleolimbot]

Thank you!

[paleolimbot]

Are there cases where we're generating WKT whose result we can't guess in advance?

[Kontinuation]

No. sd_format works with concrete ColumnarValues, so we'll always know the expected result in advance. We normalize the WKTs before comparison just for being robust.