Indexable skiplist implementations in C++

Question

All of the skip list implementations I have found so far use keys and associate them with values. But what I need is a skip list, where I can insert a value at an index position i, so that all the values following this index i can be queried with an index incremented by one.

here a little example for clarification:

//pseudocode
//let skipList sk be a list of ints, containing 5 elements.

//insert 6 at index 3
sk.insert(3, 6);

//insert 5 at index 3
sk.insert(3, 5);

//get index 4
int value = sk.get(4);

now value should be 6, because 5 was inserted at index 3 so the value 6 moved one index up.

it is possible to build a data structure like this using a skip list, see Indexable skiplist here: http://en.wikipedia.org/wiki/Skip_list

but i could not find an implementation. it would be very useful to have such a datastructure eg in situations where you have many random inserts and accesses on big lists.

Answer 1

There appears to be a link to an implementation in the wikipedia page you posted: http://code.activestate.com/recipes/576930/

It is python though.

Alternatively, take an existing C++ implementation and add the link widths to it and implement the indexed lookup.

Answer 2

I implemented a skiplist class in C# that has a method which returns the k-th largest value in the collection. It does exactly what you need for your get method.

However, the insertion method doesn't specify a position since it always keeps the list sorted therefore the position is determined by a lookup among existing values.

Take a look at https://github.com/htoma/algorithms/blob/master/Algorithms/Algorithms/Sources/SkipLists/SkipList.cs , method FindKLargestElement