Data & Analytics
dbt Fusion rejects a valid BigQuery FORMAT call with six values
Distinguish a dbt static-analysis function limit from a BigQuery SQL error.
November 9, 2026
A BigQuery model runs in the warehouse, then fails a dbt Fusion migration check. In dbt issue 17079, FORMAT with five values passes strict static analysis and the same function with six values raises dbt0209 FunctionResolutionFailed. The reporter reproduced it on dbt 2.0.11 and says the six-value SQL works in BigQuery. Google's FORMAT reference allows a format string plus repeated value arguments; it does not set the five-value ceiling shown in the dbt error.
Determine which engine rejected the model
Save the compiled SQL, dbt version, adapter, static-analysis setting and exact error. Run the smallest six-value expression in a controlled BigQuery session, then compare it with dbt compile under the same project configuration. dbt's compile reference explains the compilation step; a compile-time function-resolution failure is different from a warehouse execution failure.
Check a five-value control and a six-value case with the same data types. This avoids mistaking a type mismatch for the reported arity limit. If the model builds a business key, preserve expected output before changing the expression.
Choose a temporary migration path
The reporter gives a nested FORMAT expression as one workaround. Test that its delimiters, null handling and cast behavior produce the same key on representative data. A project-level reduction in static analysis may let valid SQL through, but it also weakens checks for unrelated models; scope and document that tradeoff.
Keep the original query and regression fixture ready for an upstream fix. The exit condition is both dbt compilation and warehouse result equivalence, not merely a green compile.
Adovo’s view
Migration errors have two possible owners: the model and the tool judging it. Prove which one failed before rewriting trusted SQL. A safe workaround preserves the business key, not just the build status.
