Add virtual `data_source` column to `columns()` for `file` and `subject` to mirror the `data_source` parameter

[abradyGDIT]

• update match_all and match_any to handle explicit requests
• reconsider allowing multiple hits to same column for multi-value columns (data_source, file_anatomic_site, ...)