Skip to main content

Spend a couple of hours investigating why a single Polyline Feature won't append to an existing FGDB feature class. It turns out because it was very long, the initial spatial index created is too small to manage the number of grids that this long polyline traverses. Never seen this before in a GDB!

 

An error occurred while flushing the cursor for the table/feature class 'LateralsGeometry'. The error number from ArcObjects is: '-2147216894'. The error message from ArcObjects is: {The spatial index grid size is invalid.}
Some error has occurred during shutdown of the writer, see previous error
Some error has occurred during shutdown of the writer, see previous error
Some error has occurred during shutdown of the writer, see previous error
FeatureWriter_4 (WriterFactory): Some error has occurred during shutdown of the writer, see previous error
Some error has occurred during shutdown of the writer, see previous error
Translation FAILED with 6 error(s) and 0 warning(s) (0 feature(s) output)

 

 

ESRI recommends vendors using their API (Eg. Safe and the GDB Writer) instead use Load Only Mode to force the GDB to recalculate the Spatial Index where necessary, but for the moment the only workaround I can think off is instead of running an Append to do a full Drop and Rebuild of the Feature Class

https://support.esri.com/en/technical-article/000010090

 

Any other suggestions?

Hi @bwn​, I don't have a better workaround unfortunately but I did come across two related resources in case you haven't already seen them:

 

  1. A short article with a bit more information on the error message,
  2. and this idea for Updating Spatial Indexes on the Esri Geodatabase writers.

 

It looks like the Idea has had some discussion that might be helpful, and I saw some hints of other users using a Python shutdown script either as an option. I've linked your question to our internal enhancement request tracked as (FMEENGINE-8678) to add an option to drop spatial indexes before writer. Unfortunately the request doesn't have a timeline just yet, but I'll post back here when the issue has been addressed. Hopefully some of these resources can provide some more ideas or connect you to other FME users who have found other workarounds.


Thanks @jovitaatsafe​ , similarly the only other workaround could think of (as referred to in these articles) was through ArcPy. Of course though that is in itself problematic as requires the FME instance to have access to the ArcPy Module, and I generally avoid having to maintain Python scripts where possible.

 

Ah well, for the moment I've just redefined the Write Mode to Drop and Create and then had to put the extra workflow data path to read in the original polyline features with the features to append and write it out as a 500,000 feature dataset (470,000 original + 30,000 to append) instead of the original workflow that was 30,000 append only.


Reply