Have been looking through all the documentation for the Data Loss Prevention service to see how we could handle a specific requirement.
There are postcode fields in a BQ table that need to be masked once the row is over 30 days old but should remain as the raw value until then
The Inspection Job has a Timescale attribute (https://cloud.google.com/dlp/docs/reference/rest/v2/InspectJobConfig#timespanconfig) that would be used to filter rows for inspection, but the De-Identification Job has no such attribute nor can it be set to run on the results of an Identification Job
Any idea how a DLP De-identification could be done on data once it reaches a certain age?
As @Samuel mentioned in the comments:
A workaround could be the use of two columns.
Each month a query updates copies the original postcode field to the new masked postcode field and deletes the original data.
So far, this is the only workaround available.
Posting the answer as community wiki since as of now, this is the only possible workaround and for the benefit of the community that might encounter this use case in the future.
Feel free to edit this answer for additional information and if there are other possible workarounds/direct solution for this use case