What causes Test Failure: expression_is_true?

A business rule constraint is violated in the source data — e.g., a negative order amount or an end date before a start date. A transformation bug inverted a calculation or applied a filter incorrectly, producing values that break the expected constraint. NULL values in one of the expression's columns cause the expression to evaluate to NULL (which is treated as false by the test). The expression references a column that changed data type after a schema migration, causing unexpected comparison behaviour. The test expression itself has a logic error — the SQL is valid but the condition is more or less restrictive than intended

How do I fix Test Failure: expression_is_true?

Run `dbt test --select --store-failures` to write failing rows to the audit schema for direct inspection.. Query the failing rows and trace the values back to the source — determine whether the bad data originated upstream or was introduced by a transformation.. Check for NULLs in the columns used in the expression: `expression_is_true` fails on NULL by default. Add a COALESCE or IS NOT NULL guard if NULLs are acceptable.. Review the expression in schema.yml and verify the SQL logic matches the intent — test it manually in a warehouse SQL editor against sample data.. If some violations are acceptable (e.g., legacy data), add a `where:` condition to the test config to exclude known exceptions.

Medium severitydata quality

Power BI Refresh Error:
Test Failure: expression_is_true

What does this error mean?

dbt's expression_is_true test (from dbt_utils) found rows where a custom SQL expression evaluated to false or NULL. This generic test allows teams to encode arbitrary business rules — like amount >= 0 or end_date >= start_date — as testable assertions in schema.yml.

Common causes

1A business rule constraint is violated in the source data — e.g., a negative order amount or an end date before a start date
2A transformation bug inverted a calculation or applied a filter incorrectly, producing values that break the expected constraint
3NULL values in one of the expression's columns cause the expression to evaluate to NULL (which is treated as false by the test)
4The expression references a column that changed data type after a schema migration, causing unexpected comparison behaviour
5The test expression itself has a logic error — the SQL is valid but the condition is more or less restrictive than intended

How to fix it

1Run `dbt test --select <test_name> --store-failures` to write failing rows to the audit schema for direct inspection.
2Query the failing rows and trace the values back to the source — determine whether the bad data originated upstream or was introduced by a transformation.
3Check for NULLs in the columns used in the expression: `expression_is_true` fails on NULL by default. Add a COALESCE or IS NOT NULL guard if NULLs are acceptable.
4Review the expression in schema.yml and verify the SQL logic matches the intent — test it manually in a warehouse SQL editor against sample data.
5If some violations are acceptable (e.g., legacy data), add a `where:` condition to the test config to exclude known exceptions.

Frequently asked questions

How do I write an expression_is_true test in schema.yml?

Under the model's `tests:` or `columns:` section: `- dbt_utils.expression_is_true: {expression: "amount >= 0"}`. For column-level tests, the column name is automatically available as the expression variable.

Can I exclude specific rows from the expression_is_true test?

Yes — add a `where:` config: `config: {where: "created_at >= '2020-01-01'"}` to restrict the test to a subset of rows.