[ARROW-16459] [C++] Update GetFileInfo in FromProto to use async filesystem APIs

• Type: Improvement
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Component/s: C++
• Labels: substrait
• External issue URL: https://github.com/apache/arrow/issues/31831
• Assignee: Unassigned
• Reporter: Ariana Villegas
• Created: 04/May/22 13:42
• Updated: 11/Jan/23 11:44

Description

GetGlobFiles function in arrrow/engine/substrait/relation_internal.cc discovery directories with sync APIs. However, it would be more efficient to use async APIs to avoid blocking calls.

for (auto res : results) {
    if (res.type() != fs::FileType::Directory) continue;
    selector.base_dir = res.path() + cur;
    ARROW_ASSIGN_OR_RAISE(auto entries, filesystem->GetFileInfo(selector));
}