Snowflake

Configuration

warehouse:
  type: snowflake
  account: myorg-myaccount     # e.g. xy12345.us-east-1
  user: KELPMESH_USER
  password: "{{ env_var('SNOWFLAKE_PASSWORD') }}"
  role: TRANSFORMER             # optional
  warehouse: COMPUTE_WH         # Snowflake virtual warehouse
  database: ANALYTICS
  schema: PUBLIC                # optional
  threads: 8

Install dependencies

pip install snowflake-connector-python

Key pair authentication (recommended)

warehouse:
  type: snowflake
  account: myorg-myaccount
  user: KELPMESH_USER
  private_key_path: /path/to/rsa_key.p8
  role: TRANSFORMER
  warehouse: COMPUTE_WH
  database: ANALYTICS

Generate a key pair:

openssl genrsa 2048 | openssl pkcs8 -topk8 -inform PEM -out rsa_key.p8 -nocrypt
openssl rsa -in rsa_key.p8 -pubout -out rsa_key.pub

Register the public key in Snowflake:

ALTER USER KELPMESH_USER SET RSA_PUBLIC_KEY='MIIBIjANBgkq...';

User permissions

CREATE ROLE KELPMESH_ROLE;
GRANT ROLE KELPMESH_ROLE TO USER KELPMESH_USER;
GRANT USAGE ON WAREHOUSE COMPUTE_WH TO ROLE KELPMESH_ROLE;
GRANT USAGE ON DATABASE ANALYTICS TO ROLE KELPMESH_ROLE;
GRANT USAGE ON SCHEMA ANALYTICS.PUBLIC TO ROLE KELPMESH_ROLE;
GRANT CREATE TABLE ON SCHEMA ANALYTICS.PUBLIC TO ROLE KELPMESH_ROLE;
GRANT CREATE VIEW ON SCHEMA ANALYTICS.PUBLIC TO ROLE KELPMESH_ROLE;
GRANT SELECT ON ALL TABLES IN SCHEMA RAW.PUBLIC TO ROLE KELPMESH_ROLE;
GRANT SELECT ON FUTURE TABLES IN SCHEMA RAW.PUBLIC TO ROLE KELPMESH_ROLE;

Materialization support

Materialization	Supported	Notes
view	✅	CREATE OR REPLACE VIEW
table	✅	CREATE TABLE AS SELECT
incremental (append)	✅	INSERT INTO
incremental (merge)	✅	Snowflake MERGE INTO ... USING
ephemeral	✅	Inlined as CTE

Incremental merge example

-- {{ config(materialized="incremental", unique_key="customer_id", incremental_strategy="merge") }}
SELECT
  customer_id,
  email,
  plan,
  updated_at
FROM raw.customers
WHERE updated_at >= CURRENT_DATE - 7

KelpMesh generates a Snowflake MERGE statement:

MERGE INTO dim_customers AS target
USING (...) AS source
ON target."customer_id" = source."customer_id"
WHEN MATCHED THEN UPDATE SET target."email" = source."email", ...
WHEN NOT MATCHED THEN INSERT ("customer_id", "email", ...) VALUES (source."customer_id", ...)

Performance tips

• Set threads: 8 or higher — Snowflake handles concurrency well.
• Use AUTO_SUSPEND on your virtual warehouse to pause it when kelpmesh runs finish.
• Tag your virtual warehouse with KelpMesh to separate compute costs from BI workloads.