Customer emails and other PII sit unmasked in a shared table

Analysts need the order data, and nobody wants raw customer emails sitting in a table everyone can query.

Intermediate to advancedRead 6 minDoing: a day8 steps

What you'll have at the end

A loaded table where raw emails and phone numbers are replaced by masked values, joins across other tables on the same email still work, and only a locked-down lookup returns the real value.

Before you start

You need

A shared table that already loads real customer emails and phone numbers next to order data, and write access to the transform step that builds it.

Not covered

Building a scanner that finds every sensitive column across a whole warehouse; this recipe scopes a masking rule to one shared table whose sensitive columns you already know.

Leans on

8 steps

Where to go next

One bad row keeps taking down the whole nightly load

go there to build the quarantine table itself; this recipe only covers making sure a raw email or phone never sits unmasked inside whatever quarantine table already exists.

Checked 25 Aug 2026

Part of the ETL (Extract, Transform, Load) cookbook