Conditionally change a field

[mbeynon]

I need to transform all rows of a table. Some fields are always transformed. One field is a message field, which can be NULL or a string such as "Hello world".

Is it possible to conditionally anonymize this field in the table, so it retains the proportion of those that were NULL vs those with a value?

Specifically, I'd like to leave all the NULL values as is, and only anonymize with a new value if it previously was a string (not NULL). The replacement would either be fixed in value such as "This is a test" or randomized by generating a few words.

The important part is making that conditional check if the original value was NULL. Can this be done with the current version of datanymizer?

Thank you.

[mbeynon]

Looks like you can do it using the Tera conditionals. Was a little unclear how to test for NULL, but this seems to work:

    template:
      # prev.message can be NULL
      format: '{% if prev.message != "\N" %}NEW VALUE{% endif %}'

[evgeniy-r]

Is it possible to conditionally anonymize this field in the table, so it retains the proportion of those that were NULL vs those with a value?

Actually, we are planning to implement something like this.