The Pan X Help describes importing of data records by appending to existing data or by replacing existing records. It is NOT about replacing fields (columns in the spreadsheet view).
I am wondering if the situation you mentioned is ideal for database use, but let’s think about your tweets example.
So I suggest you have your database open with the tweet collection in English, let’s say 50 records. You prepare your database and add new fields for the translations e.g. French and German.
You need a key field in both files that enables you to match the original tweet and the translation. This could be an additional ID number field in both files, or you simply set up your translation files in a two columns format: the original English text in the left column, the translation in the right column. Then the original tweet would be the key field.
Then you simply open the translation file as a separate database. In your original database, you select one of the empty translation columns and import the translation from the second file with the menu command Records > Morph … (using the lookup( function).
So you would be able to import a translation column into your database.