databar-table-enrichment

Databar Table-Driven Enrichment

Build a complete data enrichment pipeline: create a table, insert rows, configure an enrichment, run it on every row, and send the user a link to view results in the Databar UI.

When to use this skill

Use this when the user:

• Has a list of data (leads, companies, domains, emails) they want enriched
• Wants results stored in a persistent table they can access later
• Wants to run enrichments on a large dataset with a sharable link
• Says things like "create a table and enrich it" or "I want to enrich these rows"

If the user just wants quick results without a table, use the databar-bulk-enrichment skill instead.

Workflow

Phase 1: Prepare the data

1. Parse the user's input. Extract the list of records. Accept CSV, JSON, plain text lists, or inline data. Identify the column names.

2. Create a table. Call create_table with the table name and the columns inferred from the user's data. Save the returned table_uuid.

   create_table({ name: "Leads", columns: ["name", "company", "email"] })
   

3. Insert rows. Call create_rows with table_uuid, rows, and options: { allow_new_columns: true }.

   • Each row: { fields: { "column_name": "value", ... } }
   • Max 100 rows per request. If more than 100, split into batches and call create_rows for each batch.
   • Use allow_new_columns: true so any extra columns in the data are auto-created.

4. Verify the insert. Call get_table_columns with the table_uuid to confirm columns were created correctly.

Phase 2: Configure the enrichment

1. Find the right enrichment. Call search_enrichments with a query matching the user's goal (e.g. "email finder", "company data").

2. Inspect it. Call get_enrichment_details to see required parameters, price, and choices for any select/mselect params.

3. Resolve choices for select/mselect params. The column mapping uses table column names as values — so the enrichment runs each column value through the param. However, if a param has choices, the column values in the table must be valid choice ids.

   • choices.mode = "inline" — valid ids are in choices.items[].id
   • choices.mode = "remote" — call get_param_choices to browse valid ids
   • Warn the user if their data contains values that don't match valid choice ids.