ORC: fill initial defaults for missing id-bound fields

[cbb330]

Summary

• Build on ORC: bind struct fields by Iceberg field id in generic and Spark 3.1 readers #256's id-bound Generic ORC reader to materialize a field's initial-default only when it is absent from an id-bearing ORC file.
• Omit such fields from the read projection and fill them as a per-file constant through the reader's existing constant converter (not idToConstant).
• Preserve physical values and explicit nulls; id-less / name-mapped files retain legacy null synthesis so a default is never name-matched onto a legacy/migrated file.

Stack (bottom -> top)

1. ORC: bind struct fields by Iceberg field id in generic and Spark 3.1 readers #256 - ORC id-binding backport (base 1.2.x)
2. API: add column-default model support #260 - API column-default model (base chbush/orc-id-binding-backport)
3. this PR - ORC initial-default fill (base chbush/orc-column-default-api)
4. Spark 3.1 wiring (base chbush/orc-initial-default-orc-fill)

Depends on #260. Review this PR's own delta against its base branch chbush/orc-column-default-api (2 commits above appear until the lower PRs merge).

Supported scope

• Generic ORC row reader
• Scalar defaults at any struct nesting level, including structs within list elements and map values

Vectorized ORC and predicate pushdown on omitted default columns remain deferred.

Testing Done

• Projection tests: trusted-id omission, id-less null synthesis, required defaults, nested defaults
• Generic ORC end-to-end tests: scalar types, required/nested/list/map defaults, present values/nulls, empty/null parent structs, id-less files
• iceberg-orc and iceberg-data suites pass
• ORC RevAPI, checkstyle, and formatting pass

---

Tests generated with unit-tests plugin

Made with Cursor

[cbb330]

Closing: this was based on the wrong branch (1.2.x, which is essentially vanilla apache 1.2.x). The correct target is the OpenHouse dev branch openhouse-1.2.0, which already has the id-bound ORC readers and the NestedField column-default API (#250) — so the id-binding and API backports here were redundant.

Rebuilt as a clean 2-PR stack on openhouse-1.2.0:

• ORC: fill initial defaults for missing id-bound fields #263 - ORC: fill initial defaults for missing id-bound fields
• Spark 3.1: fill ORC initial defaults on the row reader #264 - Spark 3.1: fill ORC initial defaults on the row reader