Improve anonymize_table and anonymize_database perfs
The anon.anonymize_table
and anon.anonymize_database
functions delegates to anon.anonymize_column
which execute one SQL update per column to anonymize. This approach is valid as long as there are few columns to anonymize, but it does not scale for large datasets with many columns to anonymize.
It may be possible to refactor these two functions to perform one SQL update per table instead of one per column, which will improve performance significantly on large datasets.