Mouhssine,
Thank you for the information on the custom SQL. I'm not sure if it will help here or not. I'll provide data detail that may help to clarify what I mean.
Table Description:
id - number(20)
tin_before varchar2(20)
tin_after varchar2(20)
reason varchar2(20)
Example audit date (assume all tins are valid even though this is bogus data)
Original Data
id tin_before tin_after reason
10 applied 123-45-6789 client received tin
20 987-12-6789 applied input error
Masked Data (desired)
id tin_before tin_after
10 applied 675-54-1234
20 123-54-5678 applied
The segmentation algorithm changes the tin to a different, valid SSN but it needs to skip an alphabetic word applied.
A sql filter of:
select tin_before where tin_before != 'applied' or tin_after != 'applied';
Would skip both rows where tin_before and tin_after both need to be masked, just the word 'applied' skipped.
Does that help?
------------------------------
Lisa Garczynski
Database Engineer III
Paychex
------------------------------
Original Message:
Sent: 09-29-2020 12:43:55 PM
From: Mouhssine SAIDI
Subject: Filters and masking 2 columns
Hi Lisa,
It seems that the case isn't clear enough to give you an answer.
- can you please detail what allowed values accept your columns in clear
- what values are you expecting for masked values
- how do you set the filter, I guess you've done it from the UI
- Give an example for the before and after situation you expect.
you can set the filter using a custom sql, you can check for it here https://support.delphix.com/Delphix_Masking_Engine/Using_Custom_SQL_(KBA1764)
Regards,
Mouhssine
------------------------------
Mouhssine SAIDI
Community Member
Delphix Community Members
Original Message:
Sent: 09-29-2020 11:15:12 AM
From: Lisa Garczynski
Subject: Filters and masking 2 columns
Hello Data Masking Experts,
I have an audit table where the before/after data is on the same row to make reporting easier.
A simple example is below:
tin_audit_table
Contains 2 columns
tin_before varchar2(11)
tin_after varchar2(11)
The application allows for "applied" as a valid TIN value. I want to mask the data in both columns with a segmentation algorithm that generates valid, but different tins. If I filter the tin_before of "applied" then I'll miss the data in the tin_after that need to be masked. If I use the segmentation algorithm on "applied", I get non-conformant data and tin_after is not masked.
One solution would be two different rule sets, one for tin_before and one for tin_after that filters "applied" in either case, but that would require multiple masking jobs and filters on the same table(s).
Is there an easier way to accomplish this task? Can a segmentation algorithm be ignore alpha characters (i.e. applied)?
Thanks for your help!
------------------------------
Lisa Garczynski
Database Engineer III
Paychex
------------------------------