I am working with an adjacency matrix in a project and I need to remove an edge safely.
The project I am working on uses the matrix in the following way:
Map[Agent, Array[Edge]]. I would like to use the Array.filter function to remove the edge in question. I am doing this to avoid using indexing or zipWithIndex.
After adding four "agent"s to the matrix and adding two undirected edges this is what my matrix looks like:
A1: null, 1.1, null, null,
A2: 1.1, null, null, null,
A3: null, null, null, 1.2,
A4: null, null, 1.2, null,
As you can see the edges are (A1, A2) and (A3, A4).
To remove edge (A1, A2) I am using this:
map = map.updated(a1, map(a2).filter(e => e != null && e.toAgent != a2))
However, this is removing the null entries (as expected). This is what my matrix looks like after removing an edge using the above technique:
A1: 1.1,
A2:
A3: null, null, null, 1.2,
A4: null, null, 1.2, null,
The role of null entries is important because it allows for me to quickly insert new edges in their correct location. I am looking for a solution to filtering out a single edge.
The problem with simply doing:
map = map.updated(a1, map(a2).filter(e => e.toAgent != a2))
Is that I am operating on null entries which causes a NullPointerException (obviously). I am looking for a solution to this.
To reiterate, I want to filter using an attribute of each edge that may or may not be null. Is there a way to implement a type of "skip if null"?
Thanks for your help and advice.
You would be better off using Option rather than having null elements. Scala Option is a much more powerful and type-safe alternative to null.
A1: None, Some(1.1), None, None,
A2: Some(1.1), None, None, None,
A3: None, None, None, Some(1.2),
A4: None, None, Some(1.2), None,
You can filter an Option and it will only test Some values and skip None values.
map = map.updated(a1, map(a2).map(_.filter(_.toAgent == a2)))