dbt-superset-sync 0.1.1

Sync dbt column descriptions ↔ Apache Superset 6.x datasets

dbt-superset-sync

Sync dbt column descriptions, metrics, and labels ↔ Apache Superset 6.x datasets.

Built as a drop-in replacement for dbt-superset-lineage, with full Superset 6.x compatibility and MetricFlow support.

Features

• Push dbt model/column descriptions, verbose_name labels, and metrics → Superset datasets
• Pull published Superset dashboards → dbt exposures YAML
• Create missing Superset datasets directly from dbt models (--create-missing)
• Certify datasets with a certification badge in Superset (--certify)
• MetricFlow support (dbt 1.6+ semantic layer) alongside the legacy calculation_method format
• Column order from dbt YAML is preserved in Superset
• --no-overwrite to protect manually edited fields in Superset
• --tags, --schemas, --models filters for fine-grained control
• init command to auto-detect all settings and write .env

Installation

pip install dbt-superset-sync

Quick start

# 1. Auto-detect everything and write .env
dbt-superset-sync init

# 2. Preview what would change
dbt-superset-sync push --dry-run

# 3. Push descriptions + create missing datasets
dbt-superset-sync push --create-missing

# 4. Discover dashboards and their dbt links
dbt-superset-sync pull --list

# 5. Export selected dashboards as dbt exposures
dbt-superset-sync pull

Configuration

All options can be set via environment variables or a .env file:

SUPERSET_URL=http://localhost:8088
SUPERSET_USERNAME=admin
SUPERSET_PASSWORD=your_password

DBT_MANIFEST=/path/to/target/manifest.json
DBT_EXPOSURES_OUTPUT=/path/to/models/marts/_exposures.yml

# Auto-detected by `init`
SUPERSET_DATABASE_ID=1
SUPERSET_SYNC_SCHEMAS=gold
SUPERSET_SYNC_MODELS=dim_post,fct_account_daily
SUPERSET_SYNC_DASHBOARDS=Sales Overview,Marketing KPIs

# Owner e-mail used when creating new Superset datasets (defaults to <username>@example.com)
SUPERSET_OWNER_EMAIL=[e-mail]

Copy .env.example to get started:

cp .env.example .env

Push options

Flag	Description
--dry-run	Preview changes without writing
--create-missing	Create Superset datasets for dbt models with no matching dataset
--no-overwrite	Skip fields already filled in Superset
--certify	Mark datasets as certified
--certified-by	Certification author (default: dbt)
--certification-details	Certification details text
--continue-on-error	Continue on per-model errors instead of aborting
--schemas	Filter by dbt schema (e.g. gold)
--models	Filter by model name (e.g. dim_post,fct_account_daily)
--tags	Filter by dbt tag (e.g. marketing)
--database-id	Superset database ID for dataset creation

Column labels (verbose_name)

Add a meta.verbose_name to any dbt column — it will appear as the column label in Superset:

columns:
  - name: followers_count
    description: "Point-in-time follower count snapshot."
    meta:
      verbose_name: "Followers"

Metrics

dbt metrics (both legacy calculation_method and MetricFlow type/type_params) are pushed as Superset dataset metrics.

Supported aggregations: sum, count, average, count_distinct, ratio, derived.

Compatibility

Superset	dbt	Status
6.x	1.4 – 1.5 (legacy metrics)	✅
6.x	1.6+ (MetricFlow)	✅
< 4.x	any	❌ (use dbt-superset-lineage)

Development

git clone https://github.com/plinxore/dbt-superset-sync.git
cd dbt-superset-sync
pip install -e ".[dev]"
pytest

See CONTRIBUTING.md for pull request guidelines.

License

MIT