What ways is there to remove the sample data for some of our sensitive redshift tables?

There shd be a profiling.allow_deny_patterns.deny to block off certain columns from being profiled. Believe you can go down to table.column level

Thanks @better-orange-49102, I will try to do that. @big-carpet-38439 that would probably helpful in some cases for sure.

It would be nice to control which users could see those sample rows and/or columns.

That's a critical requirement for the feature, in other words?

I would say yes. We have Data processing agreements with some of our customers to comply with gdpr and other data privacy frameworks. The ability to 1) not profile or grab the data from those columns and/or 2) be able to pseudonymize the data before storing it in datahub would be a requirement. Disabling profiling those datasets is problematic as they would be often referenced.