For Data engineers · Updated Nov 6, 2026

Export data from Snowflake

Snowflake customers unload curated tables on a schedule to cloud storage that Scowtt can read. Your data stays under your control, and Scowtt gets read-only access to just the files you export.

How it works: You unload the agreed tables from Snowflake to cloud storage on a schedule and give Scowtt read-only access. Scowtt checks the data with you.

Where to export

Destination	Best when
A Google Cloud Storage bucket you own	Preferred. You keep full control over encryption, retention and access. See Share data through Google Cloud Storage
An Amazon S3 bucket you own	Exporting to Google Cloud Storage isn't practical. See Share data through Amazon S3
A Scowtt-managed bucket	You don't have cloud storage of your own. See Send files to a Scowtt-managed bucket

Scowtt doesn't read from Azure storage. If your Snowflake account runs on Azure, unload to Google Cloud Storage or S3.

Steps

You: Create a storage integration

Create a storage integration and external stage for your bucket, and grant Snowflake's service account permission to write there. This is separate from Scowtt's read-only access.

Snowflake's step-by-step guide for Google Cloud Storage: Configure an integration for Google Cloud Storage.
You: Unload the curated tables

Use Snowflake's COPY INTO command, for example:
```
COPY INTO @your_stage/<table>/ FROM <table> FILE_FORMAT = (TYPE = PARQUET) HEADER = TRUE
```
Without HEADER = TRUE, Parquet files lose their column names.
You: Schedule the unload

Run it with a Snowflake task, ideally hourly and at least 4 times a day, selecting only rows updated since the last run. Send the full history first.
You: Give Scowtt read access

Follow the article for your destination to grant Scowtt read-only access to the bucket or folder, then send your Scowtt team the path.
Scowtt: Check the data with you

Scowtt confirms it can read the files and validates the backfill with you before turning on recurring reads.

Using a Scowtt-managed bucket

Your Scowtt team sends the bucket path. Add it to your storage integration's allowed locations, run DESC STORAGE INTEGRATION <name>, and send Scowtt the STORAGE_GCP_SERVICE_ACCOUNT value. Scowtt grants it upload access, then you start the unload.

Good to know

Share only the curated tables Scowtt needs, in a folder used only for Scowtt.
Follow the data format standards.

Who does what

Step	Who	What changes hands
Create a storage integration	You	—
Unload the curated tables	You	History files
Schedule the unload	You	Recurring files
Give Scowtt read access	You	Read-only access and the bucket or folder path
Check the data with you	Scowtt	Confirmation, or fixes needed