Docs

Connecting Snowflake

Snowflake is planned after Databricks, tested only on simulated services. What the connection needs, what its profile covers and what is not decided.

Planned reference

Updated Oct 23, 2026

Docs menu · 25 pagesConnecting Snowflake

Status

In private development. Early access by invitation. The names, options and formats on this page can change before the first release.

On this page

Status: planned after Databricks, with no date. The Snowflake connector is built and tested only against simulated Snowflake services; it has not run on a real account, so there is no flow to follow yet. This page lists what the connection needs and what the draft Snowflake profile covers, so you can judge the fit. The Snowflake platform page has the summary.

What you need

  • A Snowflake account and a role that can read the databases and schemas you want to work with, use a warehouse, and create the objects a pipeline needs.
  • A way to sign in: key-pair authentication or a programmatic access token. Credentials are kept in your machine's credential store, as on every platform.
  • For a first try, a development database or a dedicated schema, so that nothing you test touches production.

What the profile is drafted to cover

  • Dialect: Snowflake SQL and Snowpark.
  • Table format: native Snowflake tables or Iceberg tables.
  • Merge idiom: MERGE, fed by streams, and dynamic tables.
  • Incremental tools: streams and tasks, and dynamic tables.
  • Deployment unit: SQL that creates or alters tables, streams, tasks and dynamic tables, run under your role on your warehouse, applied with the Snowflake command line, or with Terraform if your team uses it.
  • Known traps: unquoted identifiers fold to uppercase, and warehouse auto-suspend affects schedules and cost.
  • Sampling: the Snowflake connector, or COPY INTO a stage as Parquet, read into DuckDB.

The same profile fields apply to every platform; see Platform profiles.

What stays the same

The workflow is the same on every platform. You describe what you need, the agent picks a shape, fills the template for your platform and proves it with golden tests on DuckDB, and the agent deploys only after a green run. The platform tag on every generated file keeps a Databricks pipeline from deploying to Snowflake without an explicit translation step. See Deploying.

Not decided yet

  • The order in which Snowflake and Microsoft Fabric follow Databricks.
  • Which features of the Snowflake profile land first.
  • Dates. None are given.

Meanwhile

Connecting Databricks shows how the first platform connects, and the news page follows progress. If your team runs on Snowflake and would like to take part in early access, see early access.