AW: [pgrouting-users] PostGIS - Problem with driving_distancesegfault when calculating distance for many nodes

[Kurt Weninger]

I want to push this, as I am still having the problem, that the results are
wrong.
 
I was assuming that the subquery (SELECT source FROM rout_test WHERE source
> 690000) would set a focus to the nodes I am interested in.
(Additionally I sorted out costs  to other targets and added some ordering.)

Probably my assumption is wrong as the result doesn't look correct at all.

All source = target pairs do have the expected cost 0, but all other rows
have the value 1.79769313486232e+308

So I would like to ask whether my assumption is correct:

Is the function really supposed to work with a subset of nodes ?
What does the value value 1.79769313486232e+308 mean ?

BTW: I do have reverse costs, but don't have negative costs.

Trying to solve the complete matrix for all nodes leads to an out of memory
error.


I was able to solve the first problem concerning the memory issue when using
drive_destination,
so I can solve my task. Nevertheless I would like to try and verify the APSP
implementation