THE TRICKY CASE
Many times we have to mask the same column of the same table using different algorithms depending on some condition on the row being masked. In these cases we have to define one ruleset for every condition. Let's take an example: let's consider the people table; this table contains both male and female persons: the rows representing males have male first names in the column "FIRST_NAME" while the rows representing females have female first names in the column "FIRST_NAME". We have to use a specific lookup algorithm for female first names and a different specific lookup algorithm for male first names. (I work with Italian names and, at least in Italy, we use specific names depending on the gender: Mario for male and Maria for female, for example).
In order to mask the "FIRST_NAME" with two different lookup algos, with the current implementation of the masking engine, we have to define two different rulesets (every ruleset has a specific "where" condition in the table) and so we have to define two different jobs: one job for each rulset.
The disadvantage to have more jobs are:
- more complexity in launching the masking job
- more time overhead
- last, but more important: it prevents to use the provisioning masked VDB (only one job can be used)
A good idea could be to define more occurrences of the same table into the ruleset: in this way we can define different "where" condition for each table occurrence and in "Inventory" section we can select different algorithm for each table occurrence.
What do you think about this idea?