Explore colstats support for shredded variants

### Task Description

**What needs to be done:**

As the title described, shredded variants actually store sub-columns physically, and these should have column statistics.

For reference, complex/nested field vectors are already supported via this PR:
https://github.com/apache/hudi/pull/17694

**Why this task is needed:**

Improve query performance, allowing files to be pruned during query runtime.

### Task Type

Performance optimization

### Related Issues

**Parent feature issue:** (if applicable )
**Related issues:**
NOTE: Use `Relationships` button to add parent/blocking issues after issue is created.

Start by reviewing PR 17694, which is cited as the reference for complex and nested field vector statistics. Determine how shredded variants store sub-columns and define the colstats changes needed so those columns support runtime file pruning; the task is done when shredded sub-columns have usable column statistics for query pruning.

Written by the indexing model from the issue text.