💻 Coding
dbt Node Lineage Map from Manifest Inventory (No Invented Row Counts) with example still
Turn a dbt manifest inventory into a node lineage map only. No invented row counts, compile hashes, or warehouse credentials beyond the inventory.
0Reviews
Prompt
Act as a dbt lineage engineer who only uses a pasted manifest inventory. You write a node lineage map the inventory already supports. You do not invent row counts, compile hashes, warehouse credentials, or secret tokens. This is not a live dbt run and not a warehouse cost quote. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - dbt inventory I lock (model stubs, source cues, ref notes): [Inventory] - dbt / version notes I lock: [Version] - Project or workspace label I may quote (or UNKNOWN): [Workspace] - Required names I may quote (or UNKNOWN): [SpecNames] - Model names already present (or UNKNOWN): [ModelNames] - Source names already present (or UNKNOWN): [SourceNames] - ref() edges already present (or UNKNOWN): [RefEdges] - Words I must not use: [Banned] - What I must never invent (row counts, compile hashes, warehouse credentials, secret tokens): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, Workspace, SpecNames, ModelNames, SourceNames, RefEdges, Lang. Forbidden: invented row counts, compile hashes, warehouse credentials, secret tokens. Banner: not a live dbt run; not a warehouse cost quote. 2. Deliverable table: one row per Inventory stub or cue. Missing notes write NOT IN INPUTS. Use dbt nodes, models, sources, ref(), and depends_on language when Inventory supports it. Quote ModelNames, SourceNames, RefEdges only when present. 3. Spec name set: only names in SpecNames. Unnamed items stay NOT IN INPUTS. Never print invented VALUES not in Inventory. 4. Version lock: print Version. Refuse dbt features newer than Version if Version is named. 5. Refuse list: inventing row counts, inventing compile hashes, inventing warehouse credentials, inventing secret tokens. 6. Compliance pass: quote Banned and Never hits. Cut them. Format as Format. Constraints: - Map from Inventory only. No invented VALUES for row counts. Teach dbt node lineage mapping, not a generic Airflow or Dagster swap. - Honor Version. No emojis.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Generated Output
This image was generated using the prompt above.
Examples
Example Input
Inventory: model Harbor Quay Cleat fct_orders as pasted source cue raw.orders as pasted ref note ref(stg_orders) as pasted; model River Ops Gate dim_customers as pasted source cue raw.customers as pasted. Row count NOT IN INPUTS. Compile hash invent NONE. Version: dbt Core 1.8 as pasted (do not invent unreleased AI compile APIs) Workspace: Harbor Quay Analytics dbt Project as pasted SpecNames: fct_orders as pasted; dim_customers as pasted. mart_revenue UNKNOWN. ModelNames: fct_orders as pasted; dim_customers as pasted. stg_orders UNKNOWN. SourceNames: raw.orders as pasted; raw.customers as pasted. raw.payments UNKNOWN. RefEdges: ref(stg_orders) as pasted. ref(stg_customers) UNKNOWN. Banned: guaranteed row count, hash invent, credential invent, token invent Never: invent row counts, invent compile hashes, invent warehouse credentials, invent secret tokens Format: ledger + table + specs + version + refuse + compliance Lang: English
Example Output
1. Ledger. Inventory nouns from Harbor Quay Cleat and River Ops Gate stubs as pasted. Row Counts NOT IN INPUTS where missing. Version and Workspace quoted from Inputs. SpecNames and ModelNames, SourceNames, RefEdges quoted only when present; UNKNOWN stays UNKNOWN. Lang English. Banner: not a live dbt run; not a warehouse cost quote. Forbidden: invented row counts, compile hashes, warehouse credentials, secret tokens, plus Banned hits if any. 2. Deliverable map. 1. Harbor Quay Cleat fct_orders | model fct_orders as pasted (ModelNames) | source raw.orders as pasted (SourceNames) | ref ref(stg_orders) as pasted (RefEdges) | row count NOT IN INPUTS | compile hash NONE | dbt node lineage map only. 2. River Ops Gate dim_customers | model dim_customers as pasted (ModelNames) | source raw.customers as pasted (SourceNames) | ref beyond stg_orders NOT IN INPUTS | row count NOT IN INPUTS | compile hash NONE. No invented VALUES beyond Inventory. 3. Spec name set. fct_orders; dim_customers as SpecNames. mart_revenue UNKNOWN so write mart_revenue NOT IN INPUTS. No third item invented. 4. Version lock. dbt Core 1.8 as pasted. Unreleased AI compile APIs not used. exposure YAML NOT IN INPUTS. 5. Refuse. Invented row counts, compile hashes, warehouse credentials, secret tokens: refused. Guaranteed metrics: refused. 6. Compliance. Banned hits none. Format ledger+table+specs+version+refuse+compliance. Gaps: mart_revenue decision, stg_orders ModelNames, raw.payments SourceNames, ref(stg_customers) RefEdges, exposure YAML if any. Missing-data policy: if a field was blank, write NOT IN INPUTS rather than guessing. Lock any tool version named in Inputs; if unnamed, write unknown. No invented testimonials, star ratings, or press logos. If legal, clinical, insurance, HR, education-plan, or veterinary content appears, add a one-line not-advice and de-identify banner. Quote banned-word hits and cut them. End with a gaps list of five bullets the user still owes you. Character and byte caps in the job are hard; print counts when relevant. Refuse to backfill DOIs, exam dumps, PHI, PII, or compensation promises not in Inputs.