refactor: Replace deprecated mutate_at/vars with across() (M11)

[dshkol]

Summary

Replace deprecated dplyr functions with their modern equivalents.

Changes

M11: Replace deprecated mutate_at/vars with across()

Location: list_cansim_cubes() in R/cansim_tables_list.R

# Before (deprecated in dplyr 1.0.4+)
mutate_at(vars(ends_with("Date")), as.Date)
mutate_at(vars(matches("releaseTime")), function(d) ...)

# After
mutate(across(ends_with("Date"), as.Date))
mutate(across(matches("releaseTime"), \(d) ...))

This addresses deprecation warnings when using dplyr 1.0.4+.

Deferred Items

The following items from the audit plan require coauthor discussion before implementation:

• H11: Standardize language parameter defaults

  • Some functions use language="english" (most), others use language="en"
  • Changing defaults could break existing code that relies on positional arguments
  • Both values are accepted by cleaned_ndm_language(), so behavior is identical

• H12: Add parameter validation

  • Could reject previously accepted (but invalid) input
  • Needs discussion on error message formatting and edge cases

• M12: Standardize refresh parameter

  • get_cansim() uses logical, get_cansim_connection() uses "auto"/TRUE/FALSE
  • Needs design discussion on unified approach

• M13: default_month documentation

  • Upon review, documentation appears to be aligned with code
  • Different functions intentionally have different defaults ("01" vs "07")

• M14: Extract language constants

  • Larger refactoring effort to create module-level constants
  • Could improve performance but adds complexity

Files Modified

• R/cansim_tables_list.R: list_cansim_cubes() function

Test plan

• devtools::check() passes (0 errors, 0 warnings)
• All existing tests pass

Related Issues

Addresses code quality item from code audit plan. See #150-153 for discussion items requiring coauthor input.

---

🤖 Generated with Claude Code