[Feature] A `FileIO` API to list files iteratively

[smdsbz]

Search before asking

• I searched in the issues and found nothing similar.

Motivation

Currently the FileIO interface only supports listing all files / directories under a given path at a time. As a consequence callers of FileIO, e.g. ObjectRefresh, can only choose to load the entire catalog of files into memory, which may lead to poor performance and OOM.

Solution

Introduce paged list API like the following:

Pair<FileStatus[], String> listFilesPaged(
        Path path, boolean recursive, long pageSize, @Nullable String continuationToken)

This should allow implementations to take advantage of batched list APIs that are commonly seen in object stores, e.g. ListObjectsV2 with continuation token.

Anything else?

No response

Are you willing to submit a PR?

• I'm willing to submit a PR!