I was interviewed a question about implementing a bi-direction hashtable, i.e. a hashtable that stores key-value pairs such that you can find value based on key as well as find key based on value.
In another word, there is no obvious difference between a key and a value, your hashtable should have the following functions: 1. Insert(v1, v2); //Insert v1-v2 pair. 2. Find1(v); //Find v2 based on v1. Should take O(1) time. 3. Find2(v); //Find v1 based on v2. Should take O(1) time. My answer at that time was using two hashtables underlying, one treats v1 as key and v2 as value and the other treats v2 as key and v1 as value. In this way, some redundant information is kept. For example, by removing v1-v2 pair, we need to remove v1 from on hashtable and v2 from the other. I wonder whether there is alternate way that store v1-v2 pair as a whole and meanwhile allows quick search based on both v1 and v2. Many thanks. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Algorithm Geeks" group. To post to this group, send email to algogeeks@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/algogeeks -~----------~----~----~----~------~----~------~--~---
