Verify text using values you recognize
Start with names containing accents, non-Latin scripts, punctuation, line breaks and emoji. Later Jet/ACE formats use Unicode storage, while this reader treats Jet 3 text as Windows-1252. A successful parse does not validate the intended legacy code page.
If known text is wrong, stop and use an Access export with the correct source encoding. Re-encoding garbled text can make the damage permanent. Compare both short-text and long-memo columns because they may exercise different storage paths.
Understand the safer CSV representation
Spreadsheet programs can interpret text beginning with equals, plus, minus or at-sign as formulas. This exporter prefixes formula-like strings and column headers with an apostrophe in CSV. That is an intentional representation change, counted in the report.
The XLSX file uses explicit string cells without formulas or hyperlinks, and JSON retains the original text. For a database-to-database migration, prefer the typed JSON or an import process that defines field types explicitly. Do not rely on a spreadsheet’s automatic CSV inference for exact identifiers or dates.
Accept the result only after a targeted comparison
Compare every local table’s record count, then choose known rows across nulls, empty strings, exact decimals, booleans, dates, long memos and non-English text. Read every UNREAD_FIELD and UNREAD_VALUE entry. A marker means a value was not exported, not that the source contained nothing.
The browser works within fixed file, row, cell and time limits. Larger or malformed databases should be split or repaired in Access first. Your original file is never rewritten, and this tool does not promise to repair database corruption or reproduce the whole Access application.