Dynamic RLS support in Spark

OneLake Security supports Dynamic Row-Level Security (RLS), so each user sees only the rows that match their identity. This work brings Spark on Fabric into parity with that capability: Spark queries against tables protected by a Dynamic RLS role correctly return only the rows the calling user is allowed to see.With this release, Security Admins can define a single set of Dynamic RLS roles in OneLake Security and have those policies enforced uniformly across Spark, T-SQL, and Power BI. Existing Dynamic RLS roles defined in OneLake Security start working for Spark notebooks and jobs without any policy or query changes. Customers no longer have to choose between Dynamic RLS and Spark access for the same dataset.