SQL Error [XX000] with ST_AddEdgeModFace

[jeromeu]

Hello,

I am trying to manually add edges to a topology schema. I receive the following error:

SQL Error [XX000]: ERROR: Second line start point too far from first line end point

My query is

SELECT topology.ST_AddEdgeModFace('zip', 4802117, 4802118, ST_GeomFromText('LINESTRING(327015.5 6654998, 333892 6646902)', 3857))

The nodes seem correctly placed and their corredinates correspond to the entered ones for the linestring. All 4 edges connected to the nodes do have a left or right face 0.

In other words, I want to add a new edge, where there is no face yet. The nodes are already created, let's call them point B (id: 4802117, coordinates: 327015.5 6654998) on segment ABC and point E (id: 4802118, coordinates: 333892 6646902) on segment DEF that is roughly parallel to ABC. I want an edge connecting B and E.

I am using

POSTGIS="3.0.1 ec2a9aa" [EXTENSION] PGSQL="120" GEOS="3.8.0-CAPI-1.13.1 " SFCGAL="1.3.6" PROJ="Rel. 4.9.3, 15 August 2016" GDAL="GDAL 2.4.2, released 2019/06/28 GDAL_DATA not found" LIBXML="2.9.4" LIBJSON="0.12.1" LIBPROTOBUF="1.2.1" WAGYU="0.4.3 (Internal)" TOPOLOGY RASTER

and

PostgreSQL 12.1 (Ubuntu 12.1-1.pgdg18.04+1) on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 7.4.0-1ubuntu1~18.04.1) 7.4.0, 64-bit

I would deeply appreciate your help in getting around that issue.

Kind regards

[strk]

Can you provide a minimal testcase showing the problem ? We need real data (HEXEWKB format is best)

[jeromeu]

I tried, but currently I can't: When just copying the 4 edges into a new topology (or even a bit more), adding that new edge works.

Of course the topology schema I want to add that edge into is much larger than 4 edges. I ran SELECT topology.ValidateTopology, and there is a number of Errors, but none seems in the vicinity of that edge I am trying to add.

I will keep on investigating. If anyone has any idea, I am interested.

Thanks for your help

[strk]

In presence of an invalid topology (ValidteTopology returns a number of errors, you say) any bet is off. What you could do to debug this is making a copy of the topology (CopyTopology) and then reduce the size of it incrementally while still being able to reproduce the issue. Then you could enable debugging (in liblwgeom/lwgeom_topo.c and topology/postgis_topology.c there are commented-out #define on top of the file).

I'm interested in knowing if you have more info about this case.

[strk]

Closing as non-reproducible