Forum Discussion
Advocate IRequest for Case‑Insensitive Collation Support in Fabric Spark SQL for SSMS Migrations
Community SupportHi NaveenUpadhye , Thank you for reaching out to the Microsoft Community Forum.
Fabric Lakehouse notebooks and SQL endpoints run on Apache Spark semantics, where string comparisons are case sensitive by default. The spark.sql.caseSensitive configuration only controls identifier resolution (column and table names), not string value comparison. Unlike Microsoft SQL Server, Spark does not support database or column level collations and there is currently no session level option to enable case insensitive string comparison.
Currently, there is no publicly documented roadmap item indicating that SQL Server style collation behaviour will be added to Spark SQL within Microsoft Fabric Lakehouse. Collation configuration is available in Fabric Warehouse (the SQL engine), but not in the Spark engine. For large migrations, the practical patterns today are to normalize casing at ingestion or introduce compatibility views that centralize UPPER()/LOWER() logic, rather than refactoring every downstream query individually.
Your concern about migration effort and parity is valid, this is an architectural difference that can materially impact large enterprise workloads. If this gap significantly affects your migration strategy, I suggest you to raise it in the Microsoft Fabric Ideas forum so the product team can gauge demand and prioritize accordingly.
What's New? - Microsoft Fabric | Microsoft Learn
What’s new in Fabric Warehouse – July 2025 Recap | Microsoft Fabric Blog | Microsoft Fabric
What is the SQL analytics endpoint for SQL database in Fabric? - Microsoft Fabric | Microsoft Learn