ACTION_ID: snowflake_insert_rows
NAME: Snowflake: Insert Rows
CATEGORY: Integrations
CREDITS: 0

Insert a row into a Snowflake table, mapping workflow columns onto table
columns.

PREREQUISITE: Snowflake account connection — the user must create it in the Floqer UI
(Connections page); an agent cannot. Without it the action fails the row with
a connection / authentication error. Confirm it exists before configuring.

1. INPUTS

database (dynamicDropdownDatabase, required)
  Database. Name of Database to connect to.

warehouse (dynamicDropdownWarehouse, required)
  Warehouse. Name of the warehouse in Snowflake.

schema (dynamicDropdownSchema, required)
  Schema. Name of Schema to connect to.

table (dynamicDropdownTable, required)
  Table. Name of table to connect to.

role (dynamicDropdownRole, required)
  role of snowflake user

columns (dynamicJsonArray, required)
  Columns. The column names of the Table

2. OUTPUTS

(none — this action writes to the destination and returns no fields)

3. HOW TO CONFIGURE

Configure Action body:

{
  "inputs": {
    "warehouse": "COMPUTE_WH",
    "database": "ANALYTICS",
    "schema": "PUBLIC",
    "table": "ENRICHED_LEADS",
    "role": "SYSADMIN",
    "columns": [
      { "name": "EMAIL", "value": "{{person_work_email_waterfall_01ab.work_email}}" },
      { "name": "COMPANY", "value": "{{input.company_name}}" }
    ]
  }
}

Field-by-field:
  - warehouse / database / schema / table / role
                All five are required. Snowflake identifiers are
                case-sensitive when they were created quoted — match the
                casing in your account exactly.
  - columns     `[{name, value}]` against the table's column names.

4. KEY NOTES

- None of the five selector fields has a dynamic-options resolver on
  this action — Get Action Field Options returns a 400 for each. (The
  `import_from_snowflake` SOURCE does expose `warehouse` / `database` /
  `schema` / `table` / `role` resolvers; use those to discover the
  names, then paste them here.)
- `columns` likewise has no resolver. Take the column names from the
  table definition.
- Every value is sent as text. Snowflake casts on insert, so a column
  typed `NUMBER` rejects a non-numeric string — coerce upstream with a
  `formula`.
- This action always inserts. Re-running a row inserts a second copy;
  use `snowflake_upsert_rows` when the table has a natural key.
- No outputs are returned — read the cell status to tell success from
  failure.

5. WHERE IT FITS IN A WORKFLOW

Pattern: enrichment chain -> snowflake_insert_rows. Terminal step,
usually paired with a downstream warehouse job rather than more workflow
actions.

6. WHEN TO USE

Use snowflake_insert_rows to land enriched workflow rows in the
warehouse as append-only facts.

7. WHEN NOT TO USE

The table has a natural key and re-runs must not duplicate
  -> action-detail/snowflake_upsert_rows.txt

The destination is a spreadsheet
  -> action-detail/google_sheet_add_data.txt

Last updated: 2026-08-03.