Early Access Data Model Changes

This page is for early access customers who want to understand the differences between data models. Until this access summary feature moves to GA, all data model changes will be immediate. See the versioning section for more details.

See the Latest access summary export schema reference guide for the most current access summary data model.

PrPr1 PrPr2

This section describes the data model changes between prpr1 and prpr2.

You can move to the new model by simply changing your create table S3 paths to prpr2; prpr1 will no longer get files.

New tables

• datasource-metadata: This table now serves as your primary asset registry, removing technical redundancy from the tag-metadata table.

• column-metadata: This table prepares your model for the upcoming column-level access policy outcomes.

• access-requests: This table provides the primary sources for Request app access requests (who requested, what asset, what they submitted, and the request’s current/terminal status) to support access reporting and downstream audit analytics.

• request-determinations: This table provides additional insights into the determinations from access requests. Use this table with the access-requests and policy-outcomes tables to track a request from approval through access.

• request-duas: This table allows you to directly map access requests to the data use agreements the requestors are signing.

• data-products: This table provides the primary source for Request app data products to support product-level access insights.

• data-product-assets: This table maps with the data-products table to detail the exact data sources and columns within a particular data product.

Updated tables

• policy-outcomes: The column_id field has been added to policy-outcomes. It will be NULL for your current table-level data but will populate automatically as Immuta begins exporting column-level masking outcomes (Reveal policies). The scope field has also been added and represents if the merged policy is a WRITE or READ policy.

• tag-metadata: The redundancy was moved into the datasource-metadata table. data_source_id changed to target_id and target_type, which allows the tag-metadata table to directly reference both table tags and column tags. Since column IDs are strings, when joining on target_id to any other data_source_id, you will need to cast to text. Use this option as an example (dependent on your database's sql syntax):

JOIN datasource-metadata on datasource.id::TEXT = tag.target_id AND tag.target_type = 'DATA_SOURCE' 
JOIN column-metadata on column.id = tag.target_id AND tag.target_type = 'COLUMN'