Can I use Redshift's Dynamic Data Masking Feature With Fivetran?

Question

I want to mask personally identifiable information (PII) but also see the original values, so I can't use Fivetran's native column hashing feature. Can I use Redshift's Dynamic Data Masking feature with Fivetran?

Environment

Destination: Redshift

Answer

No, you cannot use Redshift's Dynamic Data Masking policies with Fivetran. This is because when updating data to Amazon Redshift, our queries use existing data to identify rows to update and optimize performance. If you mask columns in the target table, we may apply the masked data into your warehouse and corrupt your data.

If you are masking your data, you need to grant the Fivetran user read access to the actual data.