Runtime Variables

dq_persist_results

• Type: boolean
• Default: false

Controls whether history tables accumulate rows across runs.

Value	Behavior
false	History models emit zero rows. Only current-run summaries are built.
true	History models upsert current-run results, building a time series across official runs.

dq_run_id

• Type: string
• Default: dbt invocation_id

Identifier stamped on every output row. Use a meaningful value for official runs so history is easy to query and correlate.

--vars '{dq_run_id: "official_2026_03"}'

If not set, defaults to the dbt invocation ID (a UUID generated per run).

tuva_last_run

• Type: timestamp string (YYYY-MM-DD HH:MM:SS)
• Default: run_started_at

Timestamp stamped on output rows as run_ts. Use this when you want the run timestamp to reflect when the source data was produced rather than when dbt executed.

--vars '{tuva_last_run: "2026-03-01 00:00:00"}'

Examples

Current-run only (dev/test):

dbt run -s tag:spotlight

Persist to history with auto-generated metadata:

dbt run -s tag:spotlight --vars '{dq_persist_results: true}'

Persist with explicit run ID and timestamp:

dbt run -s tag:spotlight --vars '{dq_persist_results: true, dq_run_id: "official_2026_03", tuva_last_run: "2026-03-01 00:00:00"}'