thswlsqls opened a new issue, #8959: URL: https://github.com/apache/paimon/issues/8959
**Search before asking** - [x] I searched in the [issues](https://github.com/apache/paimon/issues) and found nothing similar. **Paimon version** master @ 526b91f09 (2.1-SNAPSHOT) **Compute Engine** Engine-agnostic (tables stored on Huawei Cloud OBS) **Minimal reproduce step** Read an object table (`type = object-table`) whose location is on an `obs://` path, so `org.apache.paimon.obs.HadoopCompliantFileIO` (paimon-filesystems/paimon-obs-impl) serves the listing. It does not override `FileIO.listFilesIterative(Path, boolean)`, so the default in paimon-common `fs/FileIO.java` (line 161) runs, walking the tree with one `listStatus` call per directory. **What doesn't meet your expectations?** The sibling implementations already override this method and delegate to `FileSystem.listFiles(path, recursive)`: `oss`, `s3`, `jindo` (commit `0cc45257a`), and core `HadoopFileIO`. OBS never uses the bulk listing Hadoop OBSA provides. (`gs`, `azure` and `cosn` have the same gap; this issue covers OBS only.) **Anything else?** The override keeps the returned set identical (files only). Two differences match core/oss/s3/jindo: iteration order becomes flat instead of BFS (unspecified by the contract), and `FileNotFoundException` for a missing path surfaces eagerly, not on first `hasNext()`. **Are you willing to submit a PR?** - [x] I'm willing to submit a PR! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
