You are right, S3 is a flat structure and there is no "folders". "Folders" actually just an empty object (0 bytes) just to keep a path (key) after all files with this path were deleted. You are able to upload "folders" because you are actually uploading many files with a specific key (path), but you couldn't download folder because this is not a real object, it is just a bunch of files with a same key (path). In other words to download "folder" you need to initiate download of many object as separate calls, CLI is doing it for you, but browser couldn't do that,
Thanks for your reply, Aleksandr. Like I said, I do understand that S3 is supposed to be a flat object storage space.
What I don't get is why AWS would design it like that, then re-introduce the concept of folders by treating object names with a '/' at the end differently, and then leave out half of it by not implementing downloads in some way that is at least somewhat intuitive. For example, create an archive of files with the 'folder' prefix and offer that as a download if someone marks a 'folder' and clicks the download button.
The way it currently works makes it impossible to use the S3 console as a means for sharing files, because people will start creating and uploading folders that others then can't download - inevitable chaos :(
Batch download files from multiple different folders in the same S3 bucketasked 7 months ago
disable read permission on certain folders on s3asked 3 days ago
No core, game or system folders and files missing in solution. help!Accepted Answerasked 5 years ago
Users can see ALL files and folders in search, regardless of permissionasked 3 years ago
Bulk removal of Shared Link folders and filesasked 2 years ago
S3 "folders" in console UIAccepted Answerasked 2 months ago
S3 Object deleteAccepted Answerasked 3 months ago
How to retrieve missing folders or deleted folders in Amazon Workdocs?asked 2 months ago
Renaming Empty Folders in S3
What is the best way to rename folders in AWS S3