[PostGIS] #5990: Adding 2 lines to Postgis topology takes 2 hours in a production line we have
PostGIS
trac at osgeo.org
Sun Sep 21 19:55:03 PDT 2025
#5990: Adding 2 lines to Postgis topology takes 2 hours in a production line we
have
--------------------------------+---------------------------
Reporter: Lars Aksel Opsahl | Owner: strk
Type: enhancement | Status: reopened
Priority: medium | Milestone: PostGIS 3.6.1
Component: topology | Version: 3.6.x
Resolution: | Keywords:
--------------------------------+---------------------------
Comment (by Lars Aksel Opsahl):
With no snapping (zero Tolerance in the Topology) we also get new nodes in
topology result which is correct. In this case the second line is then
added in around 4 seconds so this goes much faster.
The reason way I do not run with zero tolerance is that then I have other
cases that totally kills the performance because of all edges and faces
generated and that I have to remove afterwords.
So the main problem now is that I don't know what happens before the line
string is added and the job is done and when I get many thousands of new
edges back back after a very a long time I also have to clean up by
removing all the edges not needed in may case.
But if we had an extra parameter (max_edges) for TopoGeo_addLinestring
indicating how many edges I accept for this line and then just throw a
exception if that number is higher than max_edges , when max_edges is set.
The caller catch this exception and then decide to change tolerance or run
with higher value for max_edges for each case.
In the application I can of course try to do some kind checks upfront, but
it's better to have this checks in the Topology core code I think and then
it's just up the caller to decide if he needs to use the parameter for
max_edge or not.
--
Ticket URL: <https://trac.osgeo.org/postgis/ticket/5990#comment:9>
PostGIS <http://trac.osgeo.org/postgis/>
The PostGIS Trac is used for bug, enhancement & task tracking, a user and developer wiki, and a view into the subversion code repository of PostGIS project.
More information about the postgis-tickets
mailing list