Historical redacted data handling

1

Some of our stored data is currently redacted for integrity purposes. We would like to know whether unredacting new data entries would unredact all historical data for that particular attribute automatically? Or would the database need to be resynced or re scanned to take action to update the historical data?

Or would it require a data migration from the application code for metabase to update the historical data?

Thanks

2

Hi there!

Is this redacting happening in the database? Like, old rows have a value 'XXX' in a column but new rows have the actual value?
If that's the case, then Metabase will bring exactly what is in the database (old rows redacted, new ones unredacted).
Running a scan will improve filter options / visualizations with the new unredacted values, but no extra action would be necessary.

1 Like
3

HI Marcos, thanks for your response. So I see the old data will remain redacted as it is indeed done in the database.

Do you have redaction capabilities within Metabase itself? If so please send the docs as i cant find anything atm.

4

Check out this page in the docs. Here are some alternatives:

  • You could hide the column (Admin > Table Metadata) and disable native queries (because Metabase cannot hide anything at the SQL level). If you want to show the column’s “edited” version, you could create an SQL Model that calculates a column over the original table and build your questions starting from that model.
  • You could use data sandboxing (Pro and Enterprise) to hide or "edit" the column values that people will see (also doesn't affect native queries).
  • At the database level it's also possible (and affects native queries) but more involved, you can create a view without the column and modify Metabase's DB user's permissions to only let it see the view and not the complete table, for example.