Description:Currently, columns with BINARY or VARBINARY data types are omitted from the OAS DV dataset interface. While hiding binary payloads in table views is understandable, excluding these columns from the Join Diagram creates a significant technical limitation.
Modern high-performance databases (such as SingleStore, Snowflake, and PostgreSQL) frequently store surrogate keys, UUIDs, and MD5 hashes as BINARY or VARBINARY to optimize storage and index performance. Because these key columns are excluded from the OAS DV Join Editor, users cannot define relationships between Fact and Dimension tables that use binary keys.
Business Impact:
- Broken Semantic Layer: Star schemas relying on binary keys cannot be modeled natively in OAS DV.
- Forced Workarounds: Organizations are forced to build custom database views solely to cast binary keys to
VARCHAR, increasing database abstraction overhead and maintenance. - Incomplete Data Discovery: Users cannot perform key aggregations (such as
COUNT(DISTINCT)) on binary identifier columns.
Requested Functionality:Please enable BINARY and VARBINARY columns within the Join Diagram interface so that two entities using binary keys can be explicitly linked without requiring upstream database transformations.
