That looks like a definite bug to me. Can you send your details to the support team (or to your local reseller, if you have a support contract with one)?
Please include as much information as possible, including the full log file, the workspace, and data (if you are able) so that they can replicate the problem.
You could also mention that it might be related to FMEENGINE-63495 (in which case it would be fixed in 2021 already, and we could perhaps backport that fix to a 2020 build).
@mastertree I do not have an answer to why that issue is happening but I was able to work around the issue by simply changing the end cap style and corner style to square and miter respectively in my bufferer transformer parameters.
This issue seemed to arise with the installation of FME 2020.
@mark2atsafe I receive the same error when using the dissolver.
I added a geometryvalidator, and tried to repair it but the error happens at the same point.
I upgraded recently from 2019.2 (b19801) to 2020.0.0.1 (b20202). The workbench was created in 2019.2. It takes input (feature class) from another workbench that ran in 2020.0.0.1.
There are no upgradeable transformers up to the dissolver on which it crashes.
I can share my wokrbenches and input data
@mark2atsafe I receive the same error when using the dissolver.
I added a geometryvalidator, and tried to repair it but the error happens at the same point.
I upgraded recently from 2019.2 (b19801) to 2020.0.0.1 (b20202). The workbench was created in 2019.2. It takes input (feature class) from another workbench that ran in 2020.0.0.1.
There are no upgradeable transformers up to the dissolver on which it crashes.
I can share my wokrbenches and input data
I really can't tell what would cause that sort of message. I think the best solution is to send the details to the FME support team (or to your local reseller, if you have a support contract with one). I think that's the only way to get a resolution. I don't see that error mentioned anywhere before this.
Please include as much information as possible, including the full log file, the workspace, and data (if you are able) so that they can replicate the problem. And if we try it in 2021 and it is OK (which I suspect) then we'll know that there is a fix we can easily apply.
I had the same problem. Actually, when I upgrade Bufferer transformer it takes Arc as area parameter for end cap style and corner style. That gives an error (the same mentioned above) if I want to dissolve. The solution was for me also to change the bufferer parameters.
We encountered the bug again. It must be related to the bufferer. I can't recall, but it could very well be that using the dissolver after the bufferer creates an issue. We are trying different buffer methods to see if that avoids the crash. We will send the workbench and data to support.
We encountered the bug again. It must be related to the bufferer. I can't recall, but it could very well be that using the dissolver after the bufferer creates an issue. We are trying different buffer methods to see if that avoids the crash. We will send the workbench and data to support.
Hi @haanjohan,
yes please do if you haven't already. And as Mark mentioned, you can reference FMEENGINE-63495 which has a fix for FME 2021.0 and we can ask to see if that can be backported to FME 2020 to be made available earlier. Thanks!
Looks like using the Bufferer before the Dissolver is at issue; at least in my case. The workaround for me is now replacing the default endcaps & corners to 'Round' from 'Arc'.
The Bufferer can now 'Dissolve on Attributes' so getting rid of the Dissolver may solve the issue for most ppl. However, in my case, I cannot do that as I am adding geometry from another source and have more than one connection into the Dissolver that need not go through the Bufferer.
@mastertree I do not have an answer to why that issue is happening but I was able to work around the issue by simply changing the end cap style and corner style to square and miter respectively in my bufferer transformer parameters.
This issue seemed to arise with the installation of FME 2020.
That worked fine! Thanks. I had the same issue.
I'm getting this at the dissolver stage, no bufferer in my workspace. Solved by adding a bufferer with buffer value zero and changing end cap style to square and corner style to miter as per the workaround earlier in this thread.
I'm on FME(R) 2020.1.0.1 (20200710 - Build 20596 - WIN64) and below is the original workspace before I put in the bufferer.
I am getting the same error using the buffer transformer, the buffer just has so much unknown bugs.
also getting this same issue. AreaGapAndOverlapCleaner followed by Bufferer, change to square and miter and still have issues unfortunately. tested on 2010.2
also getting this same issue. AreaGapAndOverlapCleaner followed by Bufferer, change to square and miter and still have issues unfortunately. tested on 2010.2
Hi @oliver.morris ,
Can you try this on the latest FME 2021 beta release? It looks like a fix was released for this a few months ago.
I am getting the same error using the buffer transformer, the buffer just has so much unknown bugs.
Hi @brandonguo,
Are you getting the same error in FME 2020.1? This should have been fixed in builds b20605+ (including the 2020.2 and 2021 betas). If you are still seeing the same issue in any of the latest builds please let us know so we can re-open the reported issue.
Same issue here using 2020.1.1.1.
Also had this issue (with Bufferers leading to Dissolver.) Applied the same workaround as above (End Cap and Corner styles to Round), which allows me to get around the bug.
FME2020.1.0.1, build 20596, win64
FYI: Using FME(R) 2020.1.0.0 (20200707 - Build 20594 - WIN64)
The error Expression: edgeCurves_[circuit[0]].has_value() appears using the Dissolver on a medium
database of polygons. (103 225)
Hi, I reported the same error in FME 2020.2.0.0, Safe confirmed that it was fixed in FME 2021. I worked around the issue by buffering without the dissolving stage and adding a separate Dissolver transformer.
Exactly the same situation and error message here - Workspace built in 2019, upgraded to 2020.1, including upgrading the upgradeable transformers through FME's wizard.
Workspace containing several bufferers, including a bufferer followed immediately by a dissolver, with settings previously set to Arc (no specific reason why Arc was being used, could happily be Round).
I removed the upgraded Bufferer and Dissolver Transformers, and replaced them with new ones - with Buffer parameters set to Round instead.
Using FME 2020.1 and I am getting this error. Occurs when using 'Arc' as corner and end cap and then dissolving with a dissolver.
Using FME 2020.1 and I am getting this error. Occurs when using 'Arc' as corner and end cap and then dissolving with a dissolver.
Hi @ewanrobertmclau Yes, unfortunately this is still an outstanding issue. Please file a case and send us some sample data if you can. Meanwhile, use an ArcStroker and avoid the arc cap and corner styles in the Bufferer. We also hope to improve the error message.
I'm getting this at the dissolver stage, no bufferer in my workspace. Solved by adding a bufferer with buffer value zero and changing end cap style to square and corner style to miter as per the workaround earlier in this thread.
I'm on FME(R) 2020.1.0.1 (20200710 - Build 20596 - WIN64) and below is the original workspace before I put in the bufferer.
I'm facing the same problem of getting the edge curves error when trying to dissolve, and also do not actually need to buffer anything in my workspace. The workaround seems to be valid and the bufferer does not alter the geometry with those parameters set from what I can tell. Did you confirm that this is the case?
Cheers,
Nate
Got the same error ... checked Dissolvers in workspace and added Bufferers as mentioned - did not work for me. Hmm? Decided to go the hard way and split my dataset by half again and again and in the end found one (!) single polygon out of 10.000 which had a triple vertex what in fact caused FME to stop.
It seems this issue comes up if the input data is defect, maybe edgeCurves__edge].has_value() means there is no curve between 2 vertices.
The buffer solution may help to go around that issue but in my case it was much easier to check source geometry integrity - after deleting 3 unneeded vertices everything was fine.
Got the same error ... checked Dissolvers in workspace and added Bufferers as mentioned - did not work for me. Hmm? Decided to go the hard way and split my dataset by half again and again and in the end found one (!) single polygon out of 10.000 which had a triple vertex what in fact caused FME to stop.
It seems this issue comes up if the input data is defect, maybe edgeCurves__edge].has_value() means there is no curve between 2 vertices.
The buffer solution may help to go around that issue but in my case it was much easier to check source geometry integrity - after deleting 3 unneeded vertices everything was fine.
@hansb Could you find and repair the problem feature with a GeometryValidator?
@hansb Could you find and repair the problem feature with a GeometryValidator?
@danatsafe , I tried the GeometryValidator last year (2020.1) but it didn't solve the issue for me
