dbt-audit-helper¶

Tested version: 0.13.0 | Integration tested: Yes

dbt-audit-helper provides macros for comparing relations and queries during data model refactoring.

Dispatch configuration¶

dispatch:
  - macro_namespace: audit_helper
    search_order: ['your_project_name', 'dbt', 'audit_helper']

Macro compatibility¶

Legend: ✅ = supported on Fabric, ❌ = not supported on Fabric

Macros marked with (override) have a T-SQL-compatible override in this adapter. All other supported macros work without any adapter-specific override.

Relation and query comparison¶

Macro	Status	Notes
`compare_queries`	✅ (override)	`OFFSET/FETCH` instead of `LIMIT`; `ORDER BY` only when limit is used (T-SQL restriction)
`compare_relations`	✅ (override)	Passes `limit` through to `compare_queries` for T-SQL pagination
`compare_column_values`	✅ (override)	`CASE WHEN` instead of bare boolean expressions; adds `column_name` output, emoji toggle, custom relation names (v0.13.0)
`compare_column_values_verbose`	✅ (override)	`CASE WHEN` → 0/1 integers instead of bare boolean expressions and `coalesce(..., false)`
`compare_all_columns`	✅ (override)	Explicit `GROUP BY` instead of positional; `ORDER BY` moved out of CTE; `sum()` on 0/1 integer columns
`compare_relation_columns`	✅ (override)	Fetches metadata via `run_query()` (separate statement) to avoid distributed mode restrictions on `sys.columns`; builds comparison from `VALUES` literals
`compare_row_counts`	✅
`compare_which_query_columns_differ`	✅	Works via the adapter's `bool_or` override (`MAX(CASE WHEN ... THEN 1 ELSE 0 END)`)
`compare_which_relation_columns_differ`	✅	Delegates to `compare_which_query_columns_differ`
`quick_are_queries_identical`	✅
`quick_are_relations_identical`	✅

Not supported¶

Macro	Status	Notes
`compare_and_classify_query_results`	❌	Not dispatched; uses `true`/`false` literals (no boolean type in T-SQL)
`compare_and_classify_relation_rows`	❌	Delegates to `compare_and_classify_query_results`

Implementation notes¶

`compare_relation_columns` and `run_query()`¶

The compare_relation_columns override uses run_query() to fetch column metadata from sys.columns in a separate statement before the model is materialized. This avoids Fabric's distributed processing mode restriction that rejects sys.* queries inside CREATE TABLE AS SELECT statements. The materialized SQL only contains VALUES literals with the pre-fetched metadata.

Integration test limitations¶

The compare_relation_columns equality test is disabled in CI because the upstream package's expected-results seed CSV uses uppercase column headers (COLUMN_NAME) while the model outputs lowercase (column_name). Fabric's Latin1_General_100_BIN2_UTF8 collation makes identifiers case-sensitive, so the equality comparison fails. The macro itself works correctly — the model builds with the right data.