Protect PII in df columns from being logged

[kykyi]

Some of the dataframes I validate pandera with contain personally identifiable information (PII), things like names, email addresses, bank account numbers etc. Having these saved to logs or tools like Sentry is problematic and not best practice. Ideally this information is never exposed to such tools.

As an example, Field/Column could have a sensitive/hide_from_logs/filter_value kwarg where instead of printing [email protected] for the email col in a df, it would print out ****. This idea comes from the Rails filter_parameters config.

So if there was an invalid email in this case, instead of the failing row be printed in the error message, it would just be ****.

I am unsure of how exactly to go about implementing, but before I start I thought I would ask! I will need to implement either way, but is just a matter of if it is a PR on my private repo or here 😄

@cosmicBboy do you think it is within the domain of pandera to provide protection against this sort of thing?

Thanks!