Fix missing not_columns entries in write filters

[Copilot]

Linked issue

What changes

write filters were treating a missing not_columns name as a hard failure, even though the expected behavior is to ignore absent columns. This patch makes the recipe dataframe filter treat unknown not_columns entries as a no-op while preserving normal wildcard and regex exclusion behavior.

• Updated wrangles.recipe._filter_dataframe to skip names that do not resolve to real columns before applying exclusions.
• Kept valid not_columns selection behavior unchanged for real column names, wildcard patterns, and regex matches.
• Added a regression test covering a write recipe that includes one missing not_columns entry alongside a valid exclusion.

write:
  - dataframe:
      not_columns:
        - missing_col
        - col2

This now correctly returns the remaining columns instead of raising an error.

How it was verified

• python3 -m pytest -c pytest-local.ini tests/recipes/test_read.py tests/recipes/test_write.py -q

Compatibility and risk

Low risk. The change only relaxes the failure mode for missing not_columns entries; valid exclusions still apply, and the behavior is directly covered by focused regression tests. Safe rollback is a revert of the filter change if needed.

Ready-for-review checklist

• One human delivery owner is assigned
• The linked issue and intended milestone are correct
• The branch is current with main and has no merge conflicts
• Focused tests pass
• New or changed behavior has direct test coverage
• Documentation/schema/configuration is updated where applicable
• The PR contains no unrelated changes
• The PR description reflects the branch's current scope and latest validation
• One primary reviewer is requested only when this PR is ready

See the pull request workflow.

• Fixes not_columns should not error #1210