There's no way to do so because that's not how the system is built. Queries require indexes, which require more I/O, some level of storage consistency across th

There's no way to do so because that's not how the system is built. Queries require indexes, which require more I/O, some level of storage consistency across the system (so that the query can run against a single "snap" of the database rather than something that's constantly changing!), and a lot of other things that just aren't designed in S3. A common pattern is to store bulk data in s3 but metadata (filenames, sizes, etc.) in a database for faster queries. If you need the ability to make arbitrary queries on the data itself, S3 isn't the product to use for this. Data architecture is a fundamental part of your system; don't fuss over spending some time on it!