Using the AWS Glue catalog from another account

Apache Hive is the traditional SQL solution in the Hadoop ecosystem. Since its early versions, it has decoupled the catalog metastore from the query engine to store information about the tables and schemas. This means that multiple tools have added support to integrate with the Hive metastore over the years such as Spark, Presto/Trino, Impala, or the Hive server itself.

When AWS Glue was released, one of the value propositions that it could provide was a Hive-compatible store, which could massively scale and provide fault tolerance out of the box.

When you run EMR on EC2, you can run a Hive metastore or use the Glue catalog as the metastore just by checking the corresponding box in the cluster configuration screen (or setting the equivalent configuration if doing programmatically). When Glue is set as the catalog metastore, all the tools in the cluster that are compatible with Hive will use it to retrieve and store information about...