[NA/NA/1932 ] 'NA' value to deal with incomplete dates

A mixed date format scenario
I’m using Coda to organize scanned books, articles, historical documents and reference pictures.
Every scanned file has its own entry in a global ‘Content’ table, this means that I have to deal with mixed date formats.
Example: for books I can only record the year of publication, while for articles I have the full date. Also, sometimes part of the date is missing or not known.

My Workaround
My solution so far is to have separate columns for years, months and days.
But this is not very ideal:

  • waste of space
  • longer to update the columns
  • hard to work with in formulas

What about a new date format that accepts NA values?
Example: NA/NA/1932