Edit Question - KnowledgeHub

Title Body

I'm not sure what are prevailing opinions about using dynamic objects such as Sets as keys in Maps. I know that typical Map implementations (for example, HashMap) use a hashcode to decide what bucket to put the entry in, and that if that hashcode should change somehow (perhaps because the contents of the Set should change at all, then that could mess up the HashMap by causing the bucket to be incorrectly computed (compared to how the Set was initially inserted into the HashMap). However, if I ensure that the Set contents do not change at all, does that make this a viable option? Even so, is this approach generally considered error-prone because of the inherently volatile nature of Sets (even if precautions are taken to ensure that they are not modified)? It looks like Java allows one to designate function arguments as final; this is perhaps one minor precaution that could be taken? Do people even do stuff like this in commercial/open-source practice? (put List, Set, Map, or the like as keys in Maps?) I guess I should describe sort of what I'm trying to accomplish with this, so that the motivation will become more clear and perhaps alternate implementations could be suggested. What I am trying to accomplish is to have something of this sort: class TaggedMap<T, V> { Map<Set<T>, V> _map; Map<T, Set<Set<T>>> _keys; } ...essentially, to be able to "tag" certain data (V) with certain keys (T) and write other auxiliary functions to access/modify the data and do other fancy stuff with it (ie. return a list of all entries satisfying some criteria of keys). The function of the _keys is to serve as a sort of index, to facilitate looking up the values without having to cycle through all of _map's entries. In my case, I intend to specifically use T = String, V = Integer. Someone I talked to about this had suggested substituting a String for the Set

Tags (comma-separated) Cancel