Confirm your account has access to the Snowflake semantic views Preview feature (announced April 17, 2025) and that you are using a role with the necessary CREATE SEMANTIC VIEW privilege on the target schema
Write a CREATE OR REPLACE SEMANTIC VIEW statement specifying TABLES with optional aliases, PRIMARY KEY declarations for each logical table, and RELATIONSHIPS defining foreign-key joins between logical tables
Add a FACTS or DIMENSIONS clause to declare quantitative measures and categorical slicing columns; metrics support aggregate expressions using standard SQL aggregate functions
Optionally mark metrics or facts as PRIVATE to prevent them from being directly queried while still allowing them to be used internally in derived metric expressions
Query the semantic view using the SEMANTIC_VIEW() function in a SELECT statement, passing METRICS and DIMENSIONS arguments to retrieve governed aggregated results
Known gotchas
Snowflake semantic views are in Preview as of April 2025 and require YAML specification version 1.1 or higher with Databricks-comparable equivalence still evolving; validate behavior in non-production environments before relying on them for production dashboards
Dimensions cannot be marked as PRIVATE; only metrics and facts support the private modifier — attempting to declare a private dimension will produce a syntax error
The SEMANTIC_VIEW() query function generates SQL at runtime; complex multi-join semantic views may produce query plans that differ significantly from hand-written SQL and should be profiled for performance before production use
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp