Snowflake

Use Snowflake when that warehouse is where analysts already work. Skipprd creates the schema and tables, stages Parquet, and runs COPY INTO. You provide an account, a user, a warehouse, and a database.

Prefer a key pair. Password sign-in fails when the account enforces MFA.

Skippr, your data department

Speak to an expert. Upfront insights in days, not months.

Book an intro call →

20 minutes · No sales pitch

Before you begin

1. Account identifier in orgname-accountname form (Snowsight → account menu), for example myorg-myaccount.
2. A role with USAGE on the warehouse, database, and schema; CREATE SCHEMA on the database if Skipprd should create it; CREATE TABLE on the schema; INSERT and SELECT on the tables.
3. A key pair (recommended). Create an unencrypted PKCS#8 key and attach the public key to the user. See the Snowflake quickstart for the SQL.
4. A stage Skipprd can PUT to. The user stage @~ works. A named stage such as @SKIPPR_STAGE is fine.

export SNOWFLAKE_PRIVATE_KEY_PATH="/path/to/rsa_key.p8"

Configure

These examples assume a source named sample already exists.

Python

from skippr import Config, DataSinkSnowflake, Pipeline

cfg = Config.discover()
warehouse = cfg.data_sink(
    "warehouse",
    DataSinkSnowflake(
        account="myorg-myaccount",
        user="skippr_loader",
        private_key_path="${SNOWFLAKE_PRIVATE_KEY_PATH}",
        warehouse="COMPUTE_WH",
        database="RAW_DATA",
        schema="PUBLIC",
        role="LOADER_ROLE",
        stage="@SKIPPR_STAGE",
    ),
)
cfg.pipeline("bikehire", Pipeline(data_source=cfg.get_data_source("sample"), data_sink=warehouse))
cfg.save()

CLI

skipprd connect data-sink snowflake \
  --pipeline bikehire \
  --name warehouse \
  --account myorg-myaccount \
  --user skippr_loader \
  --private-key-path '${SNOWFLAKE_PRIVATE_KEY_PATH}' \
  --warehouse COMPUTE_WH \
  --database RAW_DATA \
  --schema PUBLIC \
  --role LOADER_ROLE \
  --stage '@SKIPPR_STAGE'

YAML

data_sinks:
  warehouse:
    Snowflake:
      account: myorg-myaccount
      user: skippr_loader
      private_key_path: ${SNOWFLAKE_PRIVATE_KEY_PATH}
      warehouse: COMPUTE_WH
      database: RAW_DATA
      schema: PUBLIC
      role: LOADER_ROLE
      stage: "@SKIPPR_STAGE"

skipprd discover --pipeline bikehire
skipprd sync --pipeline bikehire --once --log

Options

Key	Type	Required/Default	Description
account	string	Required	Account identifier (myorg-myaccount or xy12345.us-east-1)
user	string	Required	Snowflake user
warehouse	string	Required	Compute warehouse
database	string	Required	Target database
schema	string	Required	Target schema
private_key_path	path	Not set	PKCS#8 PEM for key-pair sign-in. Preferred when set.
password	secret	Not set	Password sign-in. Ignored when private_key_path is set. Use ${ENV}.
role	string	Not set	Role to assume after sign-in
stage	string	@~	Stage for uploads. Named stages such as @SKIPPR_STAGE work.
staging_uri	URI	Not set	Upload to this s3://, azure://, or gcs:// prefix instead of the Snowflake stage
staging_storage_integration	string	Not set	Snowflake storage integration. Required for GCS staging.
staging_azure_sas_token	secret	Not set	SAS token when staging_uri is Azure
staging_azure_account_key	secret	Not set	Account key when staging_uri is Azure
staging_gcs_service_account_key_path	path	Not set	Service-account JSON when staging_uri is GCS
max_concurrency	integer	Not set	Query/model only; ingest ignores it
discovery_cache_ttl_secs	integer	Not set	Query/model only; ingest ignores it

How data lands

Skipprd creates the schema and tables on the first sync, then adds columns when the source schema grows.

Each source namespace becomes one table: dots become underscores and the name is lowercased (s3.events.click_stream → s3_events_click_stream).

Each batch is staged as Parquet and loaded with COPY INTO (MATCH_BY_COLUMN_NAME = CASE_INSENSITIVE). The staged file is removed after a successful load. Retries do not duplicate rows — see Exactly-once delivery.

Skipprd type	Snowflake type
String	VARCHAR
Integer / Long	NUMBER(38,0)
Double	DOUBLE
Boolean	BOOLEAN
Date	DATE
Timestamp	TIMESTAMP_NTZ
Struct	OBJECT(...)
Array	ARRAY(...)
Map	MAP(...)

Optional staging_uri writes Parquet to your own bucket instead of the Snowflake stage. GCS staging needs staging_storage_integration because Snowflake will not take inline GCS credentials on COPY INTO.

Troubleshooting

Symptom	Fix
Failed to connect: 250001	Check account — use ORG-ACCOUNT or include the region
Incorrect username or password	Confirm user and that private_key_path or password is exported in this shell
Insufficient privileges	Grant the privileges listed above to role
390197 MFA required	Switch to key-pair auth
openssl: command not found	Install OpenSSL and recreate the key

Next steps

• Quickstart: Snowflake
• Exactly-once delivery
• skipprd sync