In case this is still relevant... It has been some time since I studied the algorithm myself, and I don't know how to execute it via a table. However, first, it is not Dijkstra's, but Prim's algorithm that finds a minimal spanning tree. The goal of Dijkstra's algorithm is to mark each vertex with the shortest distance from the initial vertex. It is not clear to me how to recover a spanning tree from the order of visited vertices.

In Dijkstra's algorithm, if you start with vertex 3, you mark vertex 2 with distance 2 and vertex 1 with distance 3, and you mark 3 as visited. Then you go to vertex 2, as having the least label, and consider its neighbor 1, but 1's label is already smaller than 2 + 2, so you leave it unchanged and mark 2 as visited. Then you go to 1, but since it has no unvisited neighbors, you mark it as visited and you are done. The shortest distance to 2 is 2 and to 1 is 3.

In Prim's algorithm, you first add an edge {2, 3}. In the next step, you could add either {1, 2} or (1, 3}, but since the former is shorter, you add it and you are done.