If you missed it ... there are no folders in S3.
The /pics/funny/cat.jpg object /pics/funny/cat.jpg not a file named cat.jpg inside a folder named funny inside another folder named pics .
In fact, this is a file with an 18-character name: pics/funny/cat.jpg . The hierarchy shown on the console is largely human-friendly, and the ability to create new folders on the console is also an illusion.
So, yes, renaming a “folder” actually means creating a new copy of each object in the “folder” with changing the names of the objects so that they look like they are on the way.
This can be done with a PUT/COPY request ($ 0.005 per 1000 depending on the region), followed by a DELETE request for the old object (free). There is no corresponding GET request, because PUT/COPY is an atomic operation inside S3, so data loading and reloading is not performed.
source share