layout
page

Building a graph and don't care.

We say a graph consists of nodes and edges,

node , edge

Edges have a source and target node, which needs to be there before you can add the edge.

Typically, when constructing a graph you need first need to specify the nodes before you can specify the edges.

Changing the order, for instance, adding y node and then the f edge from x to y will fail to produce the same graph.

Lets introducing anchors, nodes that are considered "vacant", not really there. Now we make graph construction ordering independent:

Changing the order has no effect anymore, as in, we still get the same graph in the end.

Joining datasets

What does this give us? We can now think of a union of datasets.

foo bar

Each dataset gives a graph and we join those into a bigger graph.

Suppose there is some overlapping identifier "x".

Then

gives