Skip to content

index.md

Latest commit

 

History

History
87 lines (64 loc) · 5.42 KB

index.md

File metadata and controls

87 lines (64 loc) · 5.42 KB

Package support

This adapter includes macro overrides that make popular dbt community packages work with Microsoft Fabric. Without these overrides, most community packages fail because they generate PostgreSQL-compatible SQL that neither T-SQL nor Fabric's Spark SQL accepts.

How it works

Community packages use dbt's dispatch system to allow adapters to override their macros. This adapter ships fabric__ and fabricspark__ prefixed versions of incompatible macros in its built-in macro directories.

When a package macro dispatches to find an adapter-specific implementation, dbt looks for fabric__<macro_name> (Data Warehouse) or fabricspark__<macro_name> (Lakehouse). If this adapter provides one, it takes priority over the package's default implementation.

Required dispatch configuration

To activate the macro overrides, you must add a dispatch configuration to your dbt_project.yml for each package you use. Without this, dbt's default dispatch does not search the adapter's internal macros for package-namespaced macros.

=== "Data Warehouse"

```yaml
dispatch:
  - macro_namespace: dbt_utils
    search_order: ['your_project_name', 'dbt', 'dbt_utils']
  - macro_namespace: dbt_date
    search_order: ['your_project_name', 'dbt', 'dbt_date']
  - macro_namespace: dbt_expectations
    search_order: ['your_project_name', 'dbt', 'dbt_expectations']
  - macro_namespace: dbt_external_tables
    search_order: ['your_project_name', 'dbt', 'dbt_external_tables']
  - macro_namespace: audit_helper
    search_order: ['your_project_name', 'dbt', 'audit_helper']
  - macro_namespace: dbt_profiler
    search_order: ['your_project_name', 'dbt', 'dbt_profiler']
```

=== "Lakehouse"

```yaml
dispatch:
  - macro_namespace: dbt_utils
    search_order: ['your_project_name', 'dbt', 'dbt_utils']
  - macro_namespace: dbt_date
    search_order: ['your_project_name', 'dbt', 'dbt_date']
  - macro_namespace: dbt_expectations
    search_order: ['your_project_name', 'dbt', 'dbt_expectations']
  - macro_namespace: audit_helper
    search_order: ['your_project_name', 'dbt', 'audit_helper']
  - macro_namespace: dbt_profiler
    search_order: ['your_project_name', 'dbt', 'dbt_profiler']
```

Replace your_project_name with the name field from your dbt_project.yml. Only include entries for the packages you actually use.

The dbt entry in the search order tells dbt to check the adapter's built-in macros (the "global project namespace") before falling back to the package's defaults. This is how the adapter's compatible macros take priority.

Supported packages

Package Tested version Data Warehouse Lakehouse
dbt-utils 1.3.3 ✅ Tested ✅ Tested
dbt-date 0.17.2 ✅ Tested ✅ Tested
dbt-codegen 0.14.1 ✅ Tested ✅ Tested
dbt-expectations 0.10.10 ✅ Tested ✅ Tested
dbt-audit-helper 0.13.0 ✅ Tested ✅ Tested
dbt-external-tables 0.11.0 ✅ Tested ❌ Not applicable
dbt-profiler 1.0.0 ✅ Tested ✅ Tested
dbt-project-evaluator 1.2.4 ❌ Not compatible (upstream PR) ✅ Tested

"Tested" means the adapter runs automated integration tests for that package on the given compute engine. "Via dbt-spark" means the package works through inherited dbt-spark macros without adapter-specific overrides.

!!! info "dbt-external-tables on Lakehouse"

The dbt-external-tables package is not applicable to the Lakehouse adapter. Fabric Lakehouse uses [OneLake shortcuts](https://learn.microsoft.com/fabric/onelake/onelake-shortcuts-overview?WT.mc_id=MVP_310840) for external data access, not SQL-level external tables.

Package-specific limitations

Some package features do not work due to platform limitations. The package detail pages document which specific macros or tests are affected and why.

Data Warehouse

  • dbt-expectations: regex tests (expect_column_values_to_match_regex and variants) do not work because T-SQL has no native regex support. Timeseries tests that depend on dbt-date's date arithmetic also fail.
  • dbt-audit-helper: compare_relation_columns has case-sensitivity issues with Fabric's BIN2 collation. Struct/unit test models are not supported.
  • dbt-profiler: profile_struct does not work (BigQuery-specific STRUCT type).

Lakehouse

  • dbt-expectations: regex tests produce wrong results because the upstream guards them with target.type in ['spark'] but this adapter reports fabricspark. generate_series-based tests fail with compilation errors.
  • dbt-audit-helper: struct models and compare_and_classify (which uses distinct window functions) are not supported.
  • dbt-codegen: generate_source does not work because it queries information_schema.tables, which does not exist in Spark SQL.
  • dbt-profiler: profile_struct does not work (STRUCT type is BigQuery-specific).