tycoon data transform¶
Run dbt transformations against the warehouse. Four subcommands wrap the dbt CLI.
| Command | dbt equivalent |
|---|---|
tycoon data transform run |
dbt run |
tycoon data transform test |
dbt test |
tycoon data transform build |
dbt build |
tycoon data transform docs |
dbt docs generate + dbt docs serve |
Every invocation is captured into .tycoon/metadata.duckdb (the observability DB) and re-exports the auto-generated _tycoon_dbt_* Rill dashboards.
Why not just use dbt directly?¶
You can — tycoon run dbt ... passes through. But tycoon data transform adds:
- dbt executable resolution — uses the venv-colocated
dbtso a system-widedbtfrom a different install can't shadow the version tycoon pins. - Profile resolution from
tycoon.yml— honorsdbt_profiles_dir/dbt_profile/dbt_targetfromtycoon.register dbt's persisted flags. CLI flags still win. - Observability capture — every run lands in
.tycoon/metadata.duckdbfortycoon data history. - Rill dashboard refresh — re-exports the dbt-runs Parquet so the
_tycoon_dbt_usagedashboard stays current.
Common flags¶
All four subcommands share these:
-t, --target TEXT dbt target name (dev / prod / ...). Default: tycoon.yml's
dbt_target, then dbt's own resolution
-s, --select TEXT dbt model selection syntax (e.g. 'staging+', 'tag:nightly')
--full-refresh Drop and recreate incremental models
-h, --help Show this message and exit
run — execute models¶
tycoon data transform run # all models
tycoon data transform run --select stg_widgets # one model
tycoon data transform run --select staging+ # one model + descendants
tycoon data transform run --target prod # specific target
tycoon data transform run --full-refresh # rebuild incrementals
Equivalent to dbt run plus tycoon's observability bookkeeping.
test — run dbt tests¶
tycoon data transform test # all tests
tycoon data transform test --select staging # tests for a layer
Equivalent to dbt test. Pass / fail counts surface in tycoon data history show <invocation_id>.
build — run + test together¶
tycoon data transform build # all models + tests
tycoon data transform build --select fct_orders+ # one model + descendants + their tests
Equivalent to dbt build. The recommended day-to-day command — combines run and test and stops if any test fails.
docs — generate and serve dbt docs¶
Wraps dbt docs generate followed by dbt docs serve. Press Ctrl-C to stop.
Profile / target resolution¶
For each invocation, tycoon resolves the dbt CLI args in this order:
| What | Resolution |
|---|---|
--target |
CLI flag → tycoon.yml's dbt_target → "dev" |
--profiles-dir |
tycoon.yml's dbt_profiles_dir → <dbt_project_dir>/profiles.yml if co-located → none (dbt falls back to ~/.dbt/profiles.yml) |
--profile |
tycoon.yml's dbt_profile → none (dbt uses dbt_project.yml's profile:) |
tycoon register dbt --profiles-dir / --profile / --target writes those keys into tycoon.yml so you don't pass them every invocation.
Observability captured per run¶
.tycoon/metadata.duckdb accumulates one row per invocation in dbt_runs (target, version, duration, success counts) and one row per model/test in dbt_nodes (status, exec time, rows_affected). v0.1.3+ also captures the manifest fingerprint into dbt_manifest_snapshots and a "what changed" diff into dbt_schema_changes.
tycoon data history show <invocation_id> is the terminal view over all of this.
Related¶
tycoon register dbt— setdbt_targetetc. for transform runstycoon data history— drill into a past invocation- Reference: Observability tables — full schema of
dbt_runs/dbt_nodes/dbt_manifest_snapshots - Concepts → The CLI is a thin facade over real tools