Connecting Snowflake to Streamline

Connecting a Snowflake database to your project enables Streamline to query, prefill, and write data directly to your Snowflake tables within automated workflows. This guide walks you through integrating a Snowflake data source, creating datasets, and configuring your environment for best results.

Note: This article is specific to Snowflake. If you're connecting a different data source, check our Data Source Library for platform-specific instructions.

• ML-based scanning of connected Snowflake columns for sensitive data (PII, ePHI)
• Detection categories include identity, contact, medical, financial, and location data

Automatic Data Classification:

• Update existing Snowflake records from within workflows (UPSERT)
• Insert new records when configured (requires additional Snowflake setup)
• Multiple Deliver Data steps per workflow for different destinations

Write-Back:

• Query Snowflake tables with filter operators (equals, contains, greater than, etc.)
• Prefill forms and documents with live Snowflake data
• Manual or automatic result selection in Data Search steps

Read and Search:

• Connect to any Snowflake account using private key authentication (PKCS#8)
• Browse databases, schemas, and tables through the Data Catalog
• Automatic detection of foreign key relationships for related entities

Connection Support:

Key Capabilities

Who Can Use This Feature?

Prerequisites

Before you begin, ensure the following:

• You have network access to your Snowflake account (e.g., correct region URL and access permissions).
• Your Snowflake user account has appropriate roles and privileges (e.g., USAGE and SELECT on the database, schema, and tables).
• Required firewall or network configurations are in place to allow secure outbound HTTPS traffic to Snowflake.
• Your private key is in unencrypted PKCS#8 format (see Best Practices and Requirements section below)

Tip: If you are unsure about any of these requirements, check with your database administrator or Snowflake account admin.

Part 1: Connect to a Snowflake Data Source

Step 1: Go to the Integrations Page

• From your dashboard, navigate to the Integrations tab on the left.
• Click + New Connection at the top of the page or the centre.

Step 2: Select Snowflake as Your Integration Type

• In the pop-up window, choose Snowflake from the list of supported databases.

Give your connection a recognizable name & description (e.g., Snowflake - Prod Warehouse, Snowflake Reporting).

Tip: Use names that help your team quickly understand the connection’s purpose or environment.

Step 3: Enter Snowflake Connection Details

Provide the following details:

• Host
• Warehouse
• Database (target Snowflake database)
• Username
• Private Key

Click Next to test the connection. 

Step 4: Choose Your Database and Schema

Once authenticated, you'll be prompted to select the database and schema you want to connect to.

Step 5: Finalize the Connection

Click Connect to complete the integration. Your Snowflake connection will now appear in the Integrations list.

Step 6: Explore Your Snowflake Data

Once connected:

• Navigate to the Data Catalog to browse tables and fields in your Snowflake schema.
• Use search or filters to find specific entities.
• The Entity View displays tables; the Data Field View shows individual columns with metadata (data type, nullability, etc.).

Part 2: Create and Manage a Dataset from Snowflake

Once your Snowflake connection is active, you can organize your data into reusable datasets for your projects and workflows.

Step 1: Access the Datasets Section

• Go to Datasets via the left-hand panel.
• Click + Create New Dataset.
   

Step 2: Define Dataset Basics

• Give your dataset a clear and intuitive name.
• Select your Snowflake data source.
• Choose a Primary Entity (table) from your schema.

• Add a label to make database terms more user-friendly.

  Example: Use “Sales Orders” instead of STG_ORDERS_VW.

Step 3: Configure Field Access

• Review available fields (columns).
• Configure field visibility and write access where applicable.
• Click Create to save the dataset.

Optional: Create a Dataset Directly from the Data Catalog

• From the Data Catalog, select a table.
• Click New Dataset from Entity.
• Customize the dataset by choosing fields or editing labels.

Optional: Add Related Entities

To enrich your dataset:

• During dataset creation, click Add Related Entity.
• The platform will automatically detect foreign key relationships where available (if constraints are defined in Snowflake).
• You can manually add joins as well if needed.
• Add as many related tables as needed before finalizing.

Optional: Edit an Existing Dataset

To update an existing dataset:

1. Go to Datasets and locate the dataset.
2. Click Edit to:
   • Add/remove related entities
   • Change field names or labels
   • Adjust data access settings
3. Click Next & Update to save your changes.

Troubleshooting Tips

Encountering issues? Here are some common solutions:

• Connection fails: Check account URL and warehouse/database names.
• Authentication errors: Verify that your private key is in unencrypted PKCS#8 format and that the corresponding public key is registered on the Snowflake user.
• Schema access issues: Confirm your role has USAGE and SELECT on the database and schema.
• Firewall/network problems: Ensure outbound HTTPS traffic to Snowflake is allowed.
• Tables not appearing: Only tables are discovered. If you are looking for a View or Materialized View, note that these are not currently supported.
• Insert failures: Row-by-row inserts require additional Snowflake configuration. See the "Enabling New Record Insert" companion guide.

Contact your Snowflake admin or IT support if issues persist.

Performance Tips

• Choose only necessary columns to keep datasets efficient.
• Use filters for targeted data retrieval rather than returning full tables.
• Regularly audit and update datasets to remove unused or redundant data.
• If your use case requires aggregated or transformed data, consider creating a table in Snowflake with the pre-processed results and connecting to that table.

Connecting: Best Practices and Requirements

Currently, only private key authorization is supported for connecting to Snowflake.

The private key must be in unencrypted PKCS#8 format.

Generate Your Key

Use the following command to generate a private RSA key:

openssl genpkey -algorithm RSA -out rsa_key.p8 -pkeyopt rsa_keygen_bits:2048

Extract the Public Key

Then extract the corresponding public key:

openssl rsa -in rsa_key.p8 -pubout -out rsa_key.pub

Register Public Key in Snowflake

Open a SQL Worksheet in Snowflake and run the following command, replacing your_user with your username and ...with the actual public key content:

alter user your_user set rsa_public_key='...';

Note: If you do not have the required permissions, you may need assistance from your Snowflake ACCOUNTADMIN to run this command.

Create Streamline Integration

When creating the integration in Streamline, upload your rsa_key.p8 file from your local machine to authenticate the connection.

Enabling New Record Insert from Streamline to Snowflake

Snowflake uses bulk loading by default, so enabling row-by-row inserts requires additional configuration. Follow this guide to set up Snowflake for INSERT commands from Intellistack Streamline. 

Known Limitations

• Only Snowflake tables are supported. Snowflake Views and Materialized Views are not currently discoverable through the integration and cannot be used as entities in datasets.

Supported Object Types:

Supported Data Types

Streamline currently supports a subset of Snowflake's data types. The following are supported (including their synonyms):

• VARCHAR
• TIMESTAMPTZ
• TIMESTAMPNTZ
• TIMESTAMPLTZ
• DATE
• DOUBLE
• BOOLEAN
• TIME
• NUMBER
• BINARY

For full reference, see the official Snowflake Data Types documentation.

• Only private key (PKCS#8) authentication is supported. Username/password and OAuth are not available.

Authentication:

Summary

By connecting your Snowflake environment and building datasets, you ensure your data is clean, accessible, and workflow-ready. Whether you're building datasets from scratch or using the Data Catalog, Streamline provides control over your Snowflake data.

Need help? Check our FAQs/articles or contact Support.