What data structure to use?

Sup /prog/
I've come across a problem where I need something like a hash map, but I need to be able to go both ways. That is, I should be able to retrieve a value based on a key, obviously, but I should also be able to get to the key given a value. The key set and value set contain unique values. I require logarithmic lookup time in both cases.

Is there a way to do this, without using two hashtables? I'm using sepples btw

I've come across a problem where I need something like a hash map, but I need to be able to go both ways.

http://en.wikipedia.org/wiki/Bisexual_hash_table

>>2
Wikipedia does not have an article with this exact name

Not for long...

Sure, just create an inverse of the hash function.

http://en.wikipedia.org/w/index.php?title=Talk:Bisexual_hash_table&action=history

Someone redirected it to “Bidirectional map,” then wiped it again a minute later. Second thoughts much?

Anyway, a bidirectional map is what you want. It's essentially a pair of hash tables. It's in Boost (called Bimap), so just use that. Hehe, bi.

Hehe, sepples.

DON'T HELP HIM!!!

http://en.wikipedia.org/wiki/Special:Search/Bisexual_hash_table

You searched for Bisexual hash table [Index]

Results 1-1 of 1

   * Characters of Chrono Trigger
     Relevance: 100.0% - -

what

>>9
I have always wondered how the relevance is being calculated.

>>9
>>10
article has all three words of your query

>>11
That explains only the 100% relevance.

>>12
Is that even 100% relevant, or is it just your way of explaining relevance?

^{[citation needed]}

Dude carries stone.

Hash 'em both.

Both 'em hash.

Ham Smash 'Em Up

bash 'em both

Shmup

SHIELDS UP!

>>21
NanoProbe technology

>>6
I'm not the OP, but I'm confused why you'd need anything other than a plain vanilla hash table. The requirements stated in the OP are as follows:

* Given a value, the system should be able to retrieve the corresponding key.
* Given a key, the system should be able to retrieve the corresponding value.

When using a regular associative array (implemented with a tree), I can see that you'd need a bi-directional map. With a hash table, however, you shouldn't: the first requirement is met implicitly with the hash function, and the second is synonymous with a hash lookup.

Actually, nevermind. I mis-read the OP, clearly. He wants a hash map, not a hash table.

WHATEVER LOL.

<