
If you are preparing for the Professional Data Engineer exam, Data Catalog IAM is one of those topics that looks small on paper and then turns into three or four scenario questions on test day. The exam loves to ask which role a specific persona should get when they need to tag a BigQuery table, build a tag template for the data governance team, or simply look up a metadata entry without being able to edit anything. Getting these roles straight saves you from second-guessing during the test, and it saves you from over-granting permissions when you build real catalogs in production.
In this article I want to walk through the Data Catalog roles that show up most often on the Professional Data Engineer exam, explain what each one actually lets a user do, and connect them to the least-privilege patterns Google expects you to apply.
Data Catalog is the metadata service that indexes assets across BigQuery, Pub/Sub, Cloud Storage, and other sources, and it lets you attach tags and tag templates so you can describe sensitivity, ownership, retention, and other governance attributes. Because catalog entries describe data that lives elsewhere, the IAM model is split. You have roles that control who can see and edit the metadata, and you have roles that control who can manage the structure of the tags themselves. The exam tests whether you can pick the right role for each persona without leaking access to the underlying data.
The pattern I keep coming back to is this. A user who only needs to read metadata should never get an Editor or Admin role. A user who only applies existing tags should not be able to create new tag templates. A user who designs governance schemas should manage templates without being able to delete metadata entries. Every Data Catalog question on the Professional Data Engineer exam can be answered if you internalize that split.
There are seven roles that show up in exam scenarios. I will go from narrowest to broadest, because that is how Google structures least-privilege thinking.
Policy tags are a related concept that often shows up next to Data Catalog questions, and they have their own narrower set of roles. Policy tags are how you implement column-level security in BigQuery. You define a taxonomy of policy tags, attach them to specific columns, and then grant the Fine-Grained Reader role on a tag to the users who are allowed to query that column. Users without the role see a permission error when they touch the protected column.
For the Professional Data Engineer exam, remember that policy tags are the supported pattern for restricting access to sensitive columns like email, SSN, or revenue figures. You should not try to solve column-level security with authorized views unless the question specifically rules out policy tags.
The principle of least privilege is the through-line for every Data Catalog IAM question. A few patterns that tend to be correct answers:
One more thing the exam likes to test. Data Catalog roles control access to the metadata, not to the underlying asset. A user with Data Catalog Viewer can see that a BigQuery table exists and can read its tags, but they still need BigQuery Data Viewer on the table itself to actually query it. Keep those two layers separate in your head and the scenario questions become much easier.
My Professional Data Engineer course covers Data Catalog, policy tags, and the rest of the GCP data governance surface in the depth the exam expects, with practice scenarios that map directly to the role-selection questions you will see on test day.