Serverless Synapse SQL pools enable you to read Parquet/CSV files or Cosmos DB collections and return their content as a set of rows. In some scenarios, you would need to ensure that a reader cannot access some rows in the underlying data source. This way, you are limiting the result set that will be returned to the users based on some security rules. In this scenario, called Row-level security, you would like to return a subset of data depending on the reader’s identity or role.
Row-level security is supported in dedicated SQL pools, but it is not supported in serverless pools (you can propose this feature in Azure feedback site). In some cases, you can implement your own custom row-level security rules using standard T-SQL code.
Read on to see how.