I actually versioned my database file - I had a small metadata table with version number and creation time.
Then in the output from each of my other functions, I included the database version number. So all my output could be subsequently normalized by re-running the same input versus an arbitrary version of the database file.
Then in the output from each of my other functions, I included the database version number. So all my output could be subsequently normalized by re-running the same input versus an arbitrary version of the database file.