Access S3 Tables from Databricks

[CT123456]

Hi,

When I was looking for how to access S3 Tables from Databricks, I came across this AWS blog post: https://aws.amazon.com/blogs/big-data/access-amazon-s3-iceberg-tables-from-databricks-using-aws-glue.... While it's not quite S3 Tables mentioned in the blog post (just Iceberg tables that happen to be stored in S3), the post shows a way to do it using Spark in a Databricks cluster. Does that also work for S3 Tables? Are there other ways than using Spark to access S3 Tables from Databricks?

Thanks,

C

[ashishmor]

In Databricks, you can access Iceberg tables stored in S3 using Spark, as shown in the blog. However, Databricks doesn’t currently support AWS S3 Tables directly (S3 being serverless feature from AWS). The best way is to store data in Iceberg format, register it in AWS Glue, and access it via Spark in Databricks. Other non-Spark methods aren’t supported yet on Databricks for this use case.

[CT123456]

Thanks. Any idea whether/when Databricks will support more direct, native support for accessing S3 Tables, like what Snowflake does: https://aws.amazon.com/blogs/storage/connect-snowflake-to-s3-tables-using-the-sagemaker-lakehouse-ic...

[ashishmor]

Not sure if it Databricks is supporting same way as snowflake but this is what I found (https://docs.databricks.com/aws/en/connect/storage/amazon-s3) This was updated in February 2025. Hope It helps.