Fileset is a new concept brought in 0.5.0 to manage the non-tabular data, the current implementation uses HCFS to manage the physical data. With HCFS, the Hadoop catalog should support different underlying storage, but currently we only verified local file system and HDFS.
In this issue, we should also support S3, to make the fileset hadoop catalog work with S3 object store.
Motivation
The reason to support S3 is that it is vastly used on the public cloud, we should add this support anyway.
Describe the feature
Fileset is a new concept brought in 0.5.0 to manage the non-tabular data, the current implementation uses HCFS to manage the physical data. With HCFS, the Hadoop catalog should support different underlying storage, but currently we only verified local file system and HDFS.
In this issue, we should also support S3, to make the fileset hadoop catalog work with S3 object store.
Motivation
The reason to support S3 is that it is vastly used on the public cloud, we should add this support anyway.
Describe the solution
No response
Additional context
No response