GeoJSON to GDB - Dynamic Schema Issue on Writer

Check all the coordinate systems are all set the same… and re-projected to a common grid. Perhaps you feature writer needs explicit coordinate system set on the writer parameters. You’d have to also check all the coordinate system timestamps in the log

Thanks for the reply ​@j.botterill. The coordinate system for the data features is reprojected to EPSG:2157 before the writer. Just tried to hard code it into the writer parameters too based on your suggestion, but still having the same error. Log file attached (run from FeatureHolder cache) if that helps. 
 

That error normally occurs when the schema isn’t hitting the featurewriter before the features being read. That would explain why it is intermittent, especially if you are sometimes running in feature caching mode. I normally end up doing something like this.

Thanks for the reply ​@ebygomm. The current flow I have adds an order field to the schema features and data features, so that Schema features are read first (see screenshot in the original message). I also included a FeatureHolder just before the write to make absolutely sure that schema features were present before trying to write any data. Another reason why I’m a little confused why the flow is failing on the feature type definition error.

Have you tried setting Geometry to From Schema Definition (I don’t think this should be an issue but worth a try)

Edit: I didn’t realise they hadn’t change the way the feature reader works with geojson to read multiple schemas yet

Have tried changing to Geometry from Feature Schema, but getting the same result. It looks as if it runs successfully when only 1 dataset is read at a time - so going to create an initiator workbench which runs this workbench one dataset at a time, and hopefully that will then get this working. Will let you know how it goes. Any other thoughts in why this is happening are appreciated in the meantime!

Not sure why the feature reader for geojson can’t read multiple schemas - seems a little odd.

The schema feature must arrive to the feature writer first, prior to the data feature. So the featureHolder should be on the data stream

Thanks ​@j.botterill, I tried your suggestion and had the same error when running on multiple datasets - however seems to be consistently working on 1 dataset at a time. As such, have set up an initiator workbench where datasets are chosen, then using a WorkspaceRunner to process 1 at a time - seems to be working well on Form & Flow now. Still confused as to why multiple datasets can’t be exported at the same time, but at least this is a workaround.

Thanks everyone for your help looking into this!

Hmm, thats a good approach. I also thought the Sorter might make use of Group by processing, setting it to fme_feature_type/fme_feature_type_name