Iceberg

metadata

deep-dive

The Apache Iceberg table format is an advancement beyond the original Apache Hive table format. One of the key differences is how and where metadata is stored. Hive stores its metadata in a metastore that often stores this information in a database. Iceberg takes the approach of storing metadata alongside the actual data – on the data lake itself.

This approach supports ACID-compliant transactions and creates a new version of the table each time data or structural changes occur. It allows for time-travel querying and rollbacks.

In our webinar, you will see how and where metadata is stored from the original table creation through multiple changes to the table’s contents. The presenter will utilize Starburst Galaxy for the demonstrations and you will gain access to instructions to perform the same operations yourself.