@davidl4454,
I have just run into this exact error when adding a polygonal dataset to a hosted feature layers on AGOL from an Oracle Data source.
I tracked it down to be occurring on polygons that contained more than one hole.
I added a used the 'HoleCounter' and 'Tester' to filter out any polygons that had a '_hole_count' greater than one to get to the root cause.
I am now researching is whether this is/should be allowable in the Esri Online hosted feature layer?
cheers
Steve
@davidl4454,
I have just run into this exact error when adding a polygonal dataset to a hosted feature layers on AGOL from an Oracle Data source.
I tracked it down to be occurring on polygons that contained more than one hole.
I added a used the 'HoleCounter' and 'Tester' to filter out any polygons that had a '_hole_count' greater than one to get to the root cause.
I am now researching is whether this is/should be allowable in the Esri Online hosted feature layer?
cheers
Steve
Thanks for that. It does seem to be a geometry issue. We actually have polygons on AGOL with hundreds of holes. We have had some success using a generaliser and also by deaggregating, removing very small holes, then aggregating again.
@davidl4454,
I have just run into this exact error when adding a polygonal dataset to a hosted feature layers on AGOL from an Oracle Data source.
I tracked it down to be occurring on polygons that contained more than one hole.
I added a used the 'HoleCounter' and 'Tester' to filter out any polygons that had a '_hole_count' greater than one to get to the root cause.
I am now researching is whether this is/should be allowable in the Esri Online hosted feature layer?
cheers
Steve
It looks like you are right. This is something that has only recently been happening and appears to be when the data source is a database such as Oracle in your case, or SQL in ours. Using a file geodatabase or shapefile as the source seems to overcome the problem.
It looks like you are right. This is something that has only recently been happening and appears to be when the data source is a database such as Oracle in your case, or SQL in ours. Using a file geodatabase or shapefile as the source seems to overcome the problem.
@davidl4454 Seems like you and Steve had narrowed down the cause of this error to be related to some particular geometry property, likely polygons with multiple holes.
One thing that might be worth looking into is to compare the geometry details in the source data that didn't cause the error (File GDB), with the source data that did produce the error (SQL or Oracle).
We'd definitely want to do some investigation if
- The geometries from these two sources contains identical parts, but the writer errors with one, not the other;
- OR, there are geometries that you can manually publish to AGOL but not with FME's writer.
If either of the above is the case, can you please share some sample features with the problem geometry, so we can try to reproduce the issue?
@davidl4454 Seems like you and Steve had narrowed down the cause of this error to be related to some particular geometry property, likely polygons with multiple holes.
One thing that might be worth looking into is to compare the geometry details in the source data that didn't cause the error (File GDB), with the source data that did produce the error (SQL or Oracle).
We'd definitely want to do some investigation if
- The geometries from these two sources contains identical parts, but the writer errors with one, not the other;
- OR, there are geometries that you can manually publish to AGOL but not with FME's writer.
If either of the above is the case, can you please share some sample features with the problem geometry, so we can try to reproduce the issue?
Hi Xiao Meng ?? - Thanks for looking into this one. I have attached a zip file containing two versions of a single polygon in FFS format, one written from SQL and one written from a file GDB. The SQL version fails to write to AGOL, the file GDB is successful. It appears we can publish the SQL version to AGOL manually, but it fails using FME. Please let me know if there is more I can provide.
filegdb-mssql-spatial2ffs-ffs.zip Hi Xiao Meng ?? - Thanks for looking into this one. I have attached a zip file containing two versions of a single polygon in FFS format, one written from SQL and one written from a file GDB. The SQL version fails to write to AGOL, the file GDB is successful. It appears we can publish the SQL version to AGOL manually, but it fails using FME. Please let me know if there is more I can provide.
Thank you for sharing the data, @davidl4454. I noticed that the two features are of different coordinate system. When you used FME to write to AGOL, did you use reprojection for one of these features? Perhaps the one with SQL as source was reprojected before writing? Can you send me the workspaces you used for AGOL writing? Thank you!
@davidl4454,
I have just run into this exact error when adding a polygonal dataset to a hosted feature layers on AGOL from an Oracle Data source.
I tracked it down to be occurring on polygons that contained more than one hole.
I added a used the 'HoleCounter' and 'Tester' to filter out any polygons that had a '_hole_count' greater than one to get to the root cause.
I am now researching is whether this is/should be allowable in the Esri Online hosted feature layer?
cheers
Steve
@XiaomengAtSafe Hi Xiaomeng, Please see attached. If you provide me with an email address I can give you a full run-down of our process. Thanks.
nativetitlenattestgdb.zip
arcgisonlinefeatures-mssql-spatial-filegdb2arcgiso.fmw
@davidl4454,
I have just run into this exact error when adding a polygonal dataset to a hosted feature layers on AGOL from an Oracle Data source.
I tracked it down to be occurring on polygons that contained more than one hole.
I added a used the 'HoleCounter' and 'Tester' to filter out any polygons that had a '_hole_count' greater than one to get to the root cause.
I am now researching is whether this is/should be allowable in the Esri Online hosted feature layer?
cheers
Steve
Sorry i raised the question and then went quiet.
I will send through my test data sets from oracle as well.
They were very simple 2 hole polygons that were failing though.
@XiaomengAtSafe on the coordinate system. mine are in a custom coordinate system so will test the same geometries british national grid and see if that errors as well.
Update
Good news, It appears that this issue had resolved itself yesterday (July 18th, 2018). @davidl4454 had kindly helped confirm.
@stephenneal, can you please test to see if you still see issue with your workspace? Hopefully the issue is gone for you as well. If not, that likely means the cause of your issue and David's are not quite the same. In that case, can you please file a case, and include your workspace and log file with a brief description of the issue. We will investigate further. Thank you!
Update
Good news, It appears that this issue had resolved itself yesterday (July 18th, 2018). @davidl4454 had kindly helped confirm.
@stephenneal, can you please test to see if you still see issue with your workspace? Hopefully the issue is gone for you as well. If not, that likely means the cause of your issue and David's are not quite the same. In that case, can you please file a case, and include your workspace and log file with a brief description of the issue. We will investigate further. Thank you!
Yes the issue has appeared to have been resolved as well for me!
Do we know where or what changed?
Yes the issue has appeared to have been resolved as well for me!
Do we know where or what changed?
Nothing in FME had changed, so I can only assume it might be something on AGOL side?
