Apache Iceberg™ tables: Writing data files to subdirectories in Amazon S3

Attention

This behavior change is in the 2024_06 bundle.

For the current status of the bundle, refer to Bundle History.

Snowflake writes Parquet data files in Amazon S3 for Snowflake-managed tables as follows:

Before the change:

Snowflake writes all table data files to a single directory named data/ in your external cloud storage.

Example path for each data file:

s3://externalVolumeStorageLocation/tableBaseLocation/data/snow_externalFileId.parquet

After the change:

Snowflake supports writing data files for new or existing tables to randomly-named subdirectories under the data/ directory. This helps you avoid S3 throttling and optimize query performance.

Example path for each data file:

s3://externalVolumeStorageLocation/tableBaseLocation/data/randomPrefix/snow_externalFileId.parquet

Ref: 1706