Traversing a general tree structure starting from an arbitrary node in C#

Question

I need tree traversal algorithms for arbitrary trees in both Depth-first and Breadth-first Traversal order. The tricky part is that I need to be able to start from an arbitrary node and continue until another specific node is traversed.

Now, I can use any of the ordinary algorithms and ignore the traversed nodes until I hit the start node and continue until the end node (which I currently do) but this is ugly and inefficient.

Any suggestions, please.

UPDATE: Each of my nodes have an id associated with them. In some cases, I have start and end node references to start with. In other cases, I am given two Ids, I check whether the given node is the start node or the end node by inspecting their ids. I use depth-first traversal to find the start node. Both the start and end nodes can be anywhere in the hierarchy. I hope someone could come up with an idea for the case where I am already given references to both the start-node and end-node. BTW, the nodes in the tree is actually sorted according to a sort order, which starts from 0 for each of the sub-nodes of a node and there is one root node

How would you find the start node in a tree without traversing it? — BrokenGlass, Oct 17 '11 at 12:10
Do you already *have* the node? Otherwise, you'd need a second datastructure to accelerate finding the start/end nodes. — harold, Oct 17 '11 at 12:31
Please specify how your tree is structured. Is there any sort order implemented? How are nodes related? — Christoph Grimmer, Oct 17 '11 at 12:34
he is probably searching for the node. once he finds it, he can start the DF Traversal or BF Traversal there. Are you using DFS or BFS to find the start node? — DarthVader, Oct 17 '11 at 16:34
If every node as an ID why not keep a lookup dictionary to get that node? Dictionary — Magnus, Oct 18 '11 at 13:03

score 2 · Accepted Answer · answered Oct 17 '11 at 16:41

I think what you are doing is the most efficient way to do. Especially if you are working with arbitrary trees.

You are given an arbitrary tree and you need to find the node to start the traversal from. Since there is no hierarchy (ie: it s not binary tree) you will have to scan the nodes (you might end up scanning more than half the nodes if the given node is a leaf node or if the node is not in the tree you will have to search the whole tree) until you find the node to start with. Once you find the node, you can start your DF Traversal or BF Traversal.

I dont see any optimization you can do.

score 0 · Answer 2 · answered Oct 17 '11 at 16:26

0

Rather than

Traverse(RootNode, StartNode, EndNode)

Pass the start node as the root

Traverse(StartNode, EndNode)

However, if you want to return nodes that are not children of your start node, you'll have to continue using your current method.

answered Oct 17 '11 at 16:26

Louis Ricci

20,804
5
48
62

score 0 · Answer 3 · answered Oct 18 '11 at 13:02

If you are going to have to answer lots of unrelated queries, and you need to provide a path (instead of just saying that one exists or not) then you can't do better than what you have now AFAIK.

On the other hand, if you expect lots of queries involving a small number of specific nodes (e.g. what are the paths from A to B, C, D, E, F, and G?) then you can first calculate the minimum spanning tree from that node(s) and make your BFS more efficient.

Traversing a general tree structure starting from an arbitrary node in C#

3 Answers3