How to Migrate from SQLServer to Snowflake

Why migrate from SQL Server to Snowflake?

Snowflake separates storage and compute, auto-scales, and removes index management. This reduces ops work while boosting concurrency and pay-for-what-you-use costs.

Which high-level steps are required?

1) Extract data from SQL Server.
2) Land files in cloud storage.
3) Create Snowflake objects.
4) Load data with COPY or Snowpipe.
5) Validate row counts & types.
6) Cut over applications.

How do I extract data efficiently?

Use bcp or Azure Data Factory for TB-scale exports.Export each SQL Server table to compressed CSV or Parquet to minimize network I/O.

Example bcp command

bcp dbo.Customers out customers.csv -c -t"," -S sqlserver01 -d shop -U user -P pass

How do I stage files in cloud storage?

Choose the cloud Snowflake account runs on (S3, Azure Blob, GCS). Organize folders by schema/table for parallel loads.

Example AWS CLI upload

aws s3 cp customers.csv s3://shop-migration/customers/

What Snowflake objects must be created?

Create databases, schemas, tables, and stages.Match SQL Server data types to Snowflake equivalents (e.g., VARCHAR → STRING, DATETIME → TIMESTAMP_NTZ).

DDL sample

CREATE TABLE Customers (id INT, name STRING, email STRING, created_at TIMESTAMP_NTZ);

How do I load data with COPY?

Use COPY INTO from an external or internal stage.Define file format once and reuse.

COPY example

CREATE OR REPLACE FILE FORMAT csv_ff TYPE = 'CSV' FIELD_OPTIONALLY_ENCLOSED_BY='"';CREATE OR REPLACE STAGE s3_stage URL='s3://shop-migration/' CREDENTIALS=(AWS_KEY_ID='…' AWS_SECRET_KEY='…') FILE_FORMAT=csv_ff;COPY INTO Customers FROM @s3_stage/customers/ ON_ERROR='ABORT_STATEMENT';

How do I automate continuous loads?

Configure Snowpipe to watch the storage path and auto-ingest new files. Use an AWS SNS, Azure Event Grid, or GCP Pub/Sub notification channel.

How are stored procedures migrated?

Rewrite T-SQL procedures into Snowflake JavaScript or SQL UDFs.Focus on set-based logic; avoid cursors when possible.

What validation queries should I run?

Compare row counts and checksums between SQL Server and Snowflake. Sample query:
SELECT COUNT(*), SUM(total_amount) FROM Orders;

When do I cut over the application?

After validation, switch connection strings. Keep SQL Server in read-only mode until your Snowflake workload is stable.

Best practices for large migrations?

Parallelize loads, compress files, use Snowflake multi-table COPY, and enable clustering later to avoid extra cost during ingest.

.