We are using Data Fabric Pipelines with the Upsert load method and have a question regarding Identifier Paths and key visibility.
In our table design, the UPSERT identifier is a business key, so we have implemented it as a UNIQUE constraint, not a PRIMARY KEY.
From a database perspective, the primary key is also unique, so functionally both should work.
However, we noticed the following behavior:
When the table has a PRIMARY KEY, it is visible and selectable as an Identifier Path in the Pipeline target configuration.
When the table uses a UNIQUE constraint (business key) instead, the key does not appear in the Pipeline UI as an Identifier Path.- but still upsert works.
As per standard documentation we can use either option. what is the best way ?
