Combining data level and column level matches with an 'AND' condition

  • 0
  • 1
  • Idea
  • Updated 9 months ago
  • Under Consideration
Hi All,

Currently, we can write a RegEx to match either at the data level OR at the column level, we can then combine the expressions into a Profiler Set. However, this is more like an 'OR' combine of the expressions, with the column getting flagged as sensitive if it matches one of the 2 (data level or column level).

In certain scenarios however, as an example, if i need to write a RegEx to profile the first names at the data level, the profiler also matches few other columns such as status codes, or activity codes because on a molecular level, they look like first names.

Since this issue, I have been wondering, if there could be a feature to create Profiler expressions, and then club them into groups (optional feature). This profiler group could have 2 or more expressions or groups and can be joined by an AND condition or a NOT condition, and then the expressions(which are not part of group) and groups can be combines into a profiler set, this can potentially help reduce lot of false matches and increase the flexibility offered by the profiler.

Thoughts ?

Or any other solutions to the problem i mentioned ?
Photo of Mayank Ahluwalia

Mayank Ahluwalia

  • 708 Points 500 badge 2x thumb

Posted 9 months ago

  • 0
  • 1
Photo of Gary Hallam

Gary Hallam, Official Rep

  • 2,008 Points 2k badge 2x thumb
Hi Mayank,
This sounds like a useful capability.  Check out the discussion at the link below which describes the requirement for an 'AND' capability between separate data level expressions: https://community.delphix.com/delphix/topics/exclude-identifiable-customer-numbers-from-masking.
Ultimately at the moment this needs to be a separate rule.
Regards,
Gary
Photo of Mayank Ahluwalia

Mayank Ahluwalia

  • 708 Points 500 badge 2x thumb
Thanks Gary, in masking there is still an option to achieve this by customizing the algorithm with an additional if-else, however, in profiling we don't have that option :-)
Photo of Gary Hallam

Gary Hallam, Official Rep

  • 2,008 Points 2k badge 2x thumb
Good point, I've conflated two issues here.  I will raise your suggestion.
Photo of Mayank Ahluwalia

Mayank Ahluwalia

  • 708 Points 500 badge 2x thumb
Thanks Gary
Photo of Adam

Adam

  • 140 Points 100 badge 2x thumb
I agree with Mayank's comments - this would be very usefull feature to reduce efforts to process profiling results.

Adam Przybyslawski