pgr_vidsToDMatrix - Creates a distances matrix from an array of vertex_id.


This function takes an array of vertex_id, a sql statement to select the edges, and some boolean arguments to control the behavior. It then computes kdijkstra() distances for each vertex to all the other vertices and creates a distance matrix suitable for TSP.

The function returns:

  • dmatrix:float8[] - the distance matrix suitable to pass to pgr_TSP() function.
pgr_vidsToDMatrix(IN sql text, IN vids integer[], IN directed boolean, IN has_reverse_cost boolean, IN want_symmetric boolean, OUT dmatrix double precision[])



sql:text - A SQL statement to select the edges needed for the solution.
vids:integer[] - An array of vertex_id.
directed:boolean - A flag to indicate if the graph is directed.
 boolean - A flag to indicate if the SQL has a column reverse_cost.
want_symmetric:boolean - A flag to indicate if you want a symmetric or asymmetric matrix. You will need a symmetric matrix for pgr_TSP(). If the matriix is asymmetric, the then the cell(i,j) and cell(j,i) will be set to the average of those two cells except if one or the other are -1.0 then it will take the value of the other cell. If both are negative they will be left alone.


  • kdijkstra() can fail to find a path between some of the vertex ids. We to not detect this other than the cost might get set to -1.0, so the dmatrix shoule be checked for this as it makes it invalid for TSP


  • New in version 2.1.0


This example shows how this can be used in the context of feeding the results into pgr_tsp() function. We convert a text string of x,y;x,y;... into and array of points, then convert that into an array vertex_id, then create a distance matrix that gets feed into pgr_tsp() that returns the final result.

select * from pgr_tsp(
    (select dmatrix::float8[]
       from pgr_vidstodmatrix(
                'select id, source, target, cost, reverse_cost from edge_table',
                    pgr_texttopoints('2,0;2,1;3,1;2,2;4,1;4,2;2,3;3,2', 0),
                true, true, true) as dmatrix
 seq | id
   0 |  1
   1 |  0
   2 |  6
   3 |  3
   4 |  7
   5 |  5
   6 |  4
   7 |  2
(8 rows)

This example uses the Sample Data network.

See Also