CONNECTING GOOGLE CLOUD PLATFORM(GCP) TO SNOWFLAKE: A STEP-BY-STEP GUIDE

Last updated: October 2026

Introduction

Connecting Google Cloud Platform (GCP) to Snowflake enables scalable data warehousing with automated ingestion from Google Cloud Storage. Once the connection is in place, Snowpipe with GCS Pub/Sub notifications can load new files into Snowflake tables in near real-time. This step-by-step GCP-to-Snowflake connection guide covers bucket setup, the storage integration, IAM permissions, an external stage and querying the staged files, with screenshots.

Pre-requisites

• A Snowflake account with ACCOUNTADMIN or equivalent privileges.
• A GCP account with permissions to manage storage buckets, Pub/Sub, and IAM roles.( https://console.cloud.google.com/)
• Basic familiarity with SQL and cloud storage.
• A Snowflake account hosted on GCP (or supporting cross-cloud ingestion).

Step 1: Create a Google Cloud Storage Bucket

• Log into the GCP Console and navigate to Cloud Storage > Buckets.
• Click Create bucket, enter the name snowflake2214, and select a region compatible with Snowflake (e.g., us-central1).
• Keep default settings for storage class and access control, then click Create.

Step 2: Upload a Sample File to the Folder

• In the GCP Console, go to the snowflake2214 bucket.
• Create a folder named myfolder22 by clicking Create folder.
• Navigate to myfolder22 and click Upload files to add a sample CSV file (e.g., teachers.csv) with columns: teacher_id, name, gender, subject, experience. Example content:
• teacher_id,name,gender,subject,experience
• Confirm the file appears in snowflake2214/myfolder22.

Step 3: Create a Storage Integration in Snowflake

• Open Snowflake’s web interface or SnowSQL command-line client.
• Run the following SQL to create a storage integration for GCS:

Syntax : (for creating a storage integration)

create or replace storage integration snowflake1422 --(integration name)

type = external_stage

storage_provider = gcs

enabled = true

storage_allowed_locations = ('gcs://snowflake2214/myfolder22/'); --(bucket name)

1. Retrieve the GCS service account ID

2. You can see the GCS code by running the below command.

Step 4: Grant Permissions in GCP

1. In the GCP Console, navigate to Cloud Storage > Buckets > snowflake2214.

2. Go to the Permissions tab and click Add.

3. Add the Snowflake service account from Step 3 as a principal.

4. Assign a custom role (create it first under IAM & Admin > Roles) with the storage.buckets.get, storage.objects.get and storage.objects.list permissions, which Snowflake documents for loading. Add storage.objects.delete if you will use PURGE or REMOVE, and storage.objects.create and storage.objects.delete if you will unload data. Avoid the Storage Admin role, which also lets the service account delete the bucket and change its IAM policy. The screenshot below shows Storage Admin, the role used when this guide was first written.

5. Save the changes.

Step 5: Create a Stage and File Format in Snowflake

1. Create a file format for CSV files:

2. Create an external stage pointing to the GCS folder:

3. Show the stages:

Step 6: Query the Staged Data with SELECT

1. Write the command with the identities of your table.

(ex. Teacher_id,name,section,subject…etc)

2. Run the command you’ll get your respective data which is in the google gcp.

Conclusion

Querying staged GCS files with SELECT is a quick way to check the connection and inspect the data before loading it. To load the files into a table, run COPY INTO from the same stage, and for continuous loading, set up Snowpipe with a Pub/Sub notification integration. You can also load data from Azure Blob Storage into Snowflake using a similar storage integration approach. Once data is loaded, consider using dbt for ELT data transformations to model and clean your data at scale.

Frequently Asked Questions