Anonymized dump

[carlosmuller]

Hi thanks for the super software!
What do you guys think about a option to anonymze some columns and tables?

[davidducos]

Hi @carlosmuller,
Interesting... in mydumper (export stage)? myloader (import stage)? both? How we determine the columns? filtering options? hooks?

[carlosmuller]

I am thinking in the export stage so the real data never "escape" but may be a little difficult.
I was thing in some config file with tables and columns an some options like auto generate, empty, and other generators, but for initial release could be random_unique, random, empty and fixed, something like this:

User:
   name: random
   username: random_unique
   document: empty
   createdAt: '2021-06-18'

[davidducos]

We could do something similar of what I'm doing in https://github.com/maxbube/mydumper/tree/config_file
The file content will be like:

[user]
name = random
username = random_unique
document = empty
createdAt = '2021-06-18'

[druud]

See also #84, nr 6.

[davidducos]

Phase II:
We need to create the function to load the tables from a configuration file.

[davidducos]

We could load the configuration from the sections of the default file .my.cnf but only the sections that has a . (dot) as the format will be [<schema_name>.<table_name>]. Complete match will be mandatory to identify the table. Then, each key will be the column name and the value the function name.

[davidducos]

We need to merge/complete #440 before continue with this

[davidducos]

@carlosmuller another phase has been completed.

[carlosmuller]

@davidducos thanks will take a look as soon I have time