Address List Structure in HBase

Question

I am trying to implement the following graph reduction algorithm in

• Count - undirected weighted graph
• I want to delete all nodes with only two neighbors
• and update the scales

Take a look at the following illustration:

Graph Reduction Algorithm http://public.kungi.org/graph-reduction.png

The algorithm should convert the top graph to the bottom. Eliminate node 2 and update the edge weight to: w (1-3) = w (1-2) + w (2-3)

Since I have a very large chart, I do this with MapReduce.

My question is how to present a graph in HBase. I was thinking of creating an adjacency list structure in HBase as follows:

Column family: nodes, neighbors 1 → 2, 6, 7 ...

Is there a better way to do this?

Answer 1

The most recommended structure is the list of adjacency.

You can use each node identifier as a row identifier and neighbor identifiers as column qualifiers, with weights as values.