Idea from pgantomizer - load dump into local dev db, anonymize it on the fly without parsing

[zealot128]

Thanks for your talk yesterday!

I just found this Python tool which does something similar goal, like your tool https://github.com/asgeirrr/pgantomizer/tree/master/pgantomizer

But their approach is more like your First Apporch "Read + Update", I guess?

• They can load a dump.sql into a (local) PG database, connect with it, and than run queries to extract table information, etc. and generate a anonymized version with it - instead of working on the dump.

Their config format also has some more options, which might be also something we discussed for your main Masking Gem yesterday

• each column has several options:
  • aggregate_length - replaces content of the column with its length (can be used on any type that supports length function)
  • clear - simply nulls out the value (whatever DB constraints still apply)
  • example_email - replaces the value with an @example.com based on the primary key value
  • md5 - alternative to default TEXT handling, useful for creating variance aside default handling while also guaranteeing value uniqueness
  • x_out - converts a string alpha-numeric characters to X's, retaining length
• truncate table - Great idea! Like analytics/usage data might not be relevant in some scenarios in dev

[kibitan]

@zealot128 thank you for letting me know! and i'm happy to hear that you liked the talk :) interesting project, yes i also guess it is "Read + Update" approach. and as you suggest configuration option looks interesting, i would closely look later! thanks!!