Opened 7 years ago
Closed 7 years ago
#6876 closed defect (wontfix)
GNM failed to connect nodes and edges
| Reported by: | alexbruy | Owned by: | bishop |
|---|---|---|---|
| Priority: | normal | Milestone: | |
| Component: | GNM | Version: | svn-trunk |
| Severity: | normal | Keywords: | gnm |
| Cc: |
Description
I noticed that GNM failed to connect nodes and edges on some datasets using ConnectPointsByLines and as result it is not possible to find shortest path(s).
Tested on Linux and Windows with self-compiled GDAL (2.1.2, 2.2.0 and trunk). I tried to use layers in WGS84 and in projected CRS, and different tolerance values.
Sample dataset and script attached.
Attachments (1)
Change History (4)
by , 7 years ago
| Attachment: | gnm-test.tar.xz added |
|---|
comment:1 by , 7 years ago
| Component: | Algorithms → GNM |
|---|---|
| Owner: | changed from to |
comment:2 by , 7 years ago
| Status: | new → assigned |
|---|
comment:3 by , 7 years ago
| Resolution: | → wontfix |
|---|---|
| Status: | assigned → closed |
This is a problem with your data. Now in GNM the connect nodes functions works such way: search lines begins and ends and points from nodes layer. If they all are in tolerance - the new connection added to graph. The points not divides lines in segments now. Expected that you have to prepare data manually or in script. The option to divide lines into segments in GNM are the subject to discuss in mailing list.
By the way, there are several nodes which are corespondents to the current GNM behaviour. For example: 29 and 263
gnmanalyse dijkstra 29 263 ~/tmp/gnmnetwork
Layer name: Dijkstra
Geometry: Geometry Collection
Feature Count: 9
Here the screenshot of problem (Node 126 is not at the begin or end of the line)
Test data and script