without an extension, the object is handled as CSV. and since the Parquet object is binary, it probably failed upon trying to read the object.
this behavior was changed in PR #49411 (merged) RGW identified the input type according to AWS-CLI parameters and not according to object extension.
this issue resolved in https://github.com/ceph/ceph/pull/49411/
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Red Hat Ceph Storage 7.0 Bug Fix update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2023:7780