Replace anonymized data in string

On 12 Nov 2021, at 15:12, Patrick FICHE <Patrick.Fiche@aqsacom.com> wrote:

Is there any way to do this ?

There was a presentation on the subject of anonymization and data masking at
Fosdem PGDay 2019, maybe the slides from there can give any insights?

https://www.postgresql.eu/events/fosdem2019/schedule/session/2287-anonymization-and-data-masking-with-postgresql/

--
Daniel Gustafsson https://vmware.com/

On 11/12/21 7:12 AM, Patrick FICHE wrote:

Hi Team,

I have some data that has been anonymized and I would like to generate
some test data from this. In some way, I would like to deanonymize
this data with random data.

For example, phone numbers have been anonymized with changing the 5
right digits with the 8 digit (preserving length).

Applying this, the number 390694802756 was changed to 3906948088888.

I would like to get random digits at the end of the phone number
knowing that anonymized data can be a variable length.

So, I would like to change every sequence (at least 2) of 8 by random
value of same length (I don�t worry if phone number contains 88 in the
middle and this sequence is changed to random data)�

I tried to do this with replace / regexp_replace functions but could
not achieve what I wanted to do.

I don�t want these digits to be changed with a single one (88888 by
111111 or 99999 but something like 42384)�

Ideally, the new string would be different if multiple sequences of 8
appear in a string and would be different from one record to the other
when applied to a full table�

Is there any way to do this ?

Thanks,

Patrick

Usual trick it to select floor(random()*100000);

On 12 Nov 2021, at 15:12, Patrick FICHE <Patrick.Fiche@aqsacom.com> wrote:

Is there any way to do this ?

There was a presentation on the subject of anonymization and data masking at Fosdem PGDay 2019, maybe the slides from there can give any insights?

https://www.postgresql.eu/events/fosdem2019/schedule/session/2287-anonymization-and-data-masking-with-postgresql/

--

Thanks a lot for your answer.
This is a very good presentation for anonymization technics.
Unfortunately, my data has been already anonymized and I'm trying to random the anonymized part... which is a bit different from what I could find here 😊