Timeline for How does pgRouting shortest_path work?
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 15, 2012 at 11:26 | history | tweeted | twitter.com/#!/StackGIS/status/158510368455532544 | ||
| Jan 10, 2012 at 17:08 | comment | added | underdark | You could run your old vector file through grass v.clean to get a version with split lines at every intersection. | |
| Jan 10, 2012 at 16:17 | vote | accept | Antony | ||
| Jan 10, 2012 at 16:04 | comment | added | Antony | I can tell you what does not work for sure, if I digitise 10 parallel lines and then intersect all of them with one line, these intersections are not identified when I call assign_vertex_id. Now, if I digitise a line and then start a new line snaped at the end of the previous, this vertex is identified. But of course I want the first technique to work as well..So this probably means I have to digitise smaller lines and connect them using the end of the previous line as the beginning of the next? That would create lots of unnecessary data...but would work | |
| Jan 10, 2012 at 15:27 | comment | added | mrg | Do you have a tool like QGIS? If so, load your network with that and put the layer in Edit mode. That way you see all verices and you can probably spot missing vertices. Otherwise select vertices that are only 1 time a source or target and make them visible to see where they lie. Dead ends are ok, but somewhere inside your network is probably wrong. | |
| Jan 10, 2012 at 15:08 | comment | added | Antony | Sure, I tested what you suggested and it works with specific examples but not with others, so I guess I have wrongly connected vertices. So in my db there are sources and targets like this (gid:1,s:1,t:2),(gid:2,s:3,t:4)...In this case though, gid:1 is intersecting gid:2 but I guess a vertex is not properly created, eitherwise it would be picking it up.. | |
| Jan 10, 2012 at 15:00 | answer | added | Mapperz♦ | timeline score: 4 | |
| Jan 10, 2012 at 14:53 | comment | added | mrg | Start with a very simple case. The simplest test would be to route from a source to target on the same edge. If that doesn't work you might not have to correct columns. If it works test a target that is on a connecting edge. If that doesn't work your edges might not be correctedly conected, etc. | |
| Jan 10, 2012 at 14:17 | history | edited | underdark | CC BY-SA 3.0 |
edited title
|
| Jan 10, 2012 at 13:58 | answer | added | underdark | timeline score: 3 | |
| Jan 10, 2012 at 13:51 | history | edited | underdark | CC BY-SA 3.0 |
edited title
|
| Jan 10, 2012 at 12:54 | history | asked | Antony | CC BY-SA 3.0 |