If you do not want to use a query of the form or n1ql, there is no way to get documents without knowing their exact keys. That is, you can only get keys based on the prefix, if you have a way to pre-create possible keys on the client side, for example. User-1, User-2 ... User-n.
However, you can query the prefix you are talking about in n1ql without creating any additional indexes, because with n1ql you will already have the main index for all the keys in the document. So you can do something like "SELECT META(myBucket).id FROM myBucket WHERE META(myBucket).id LIKE "prefix%";
source share