Question

Why do file geodatabase and SDE readers produce different fme_geometry types?

  • 14 January 2013
  • 8 replies
  • 2 views

Badge

  I have one very simple workbench that I am running against both SDE (9.3) and FGDB (10.0) sources.

 

  I noticed two different fme_geometry values for same type of objects.

 

  For the first one, i.e SDE 9.3, I have 

fme_geometry=fme_donut
, but for the same features in FGDB 10.0 I have 
fme_geometry=fme_aggregate

  Is this a "problem" of converting ESRI data from 9.3 to 10.0 or how FME interprets geometry?

 

  Because in case of fme_aggregate, FME is doing "additional" job of constructing donuts again:

 
 Calculating inside points Determining uppermost containment Constructing donuts from xxx points 
 

  (Visible in log window)


8 replies

Badge +21
Both of them should support fme_aggregate. However it seems that the SDE 9.3 was not stored as fme_aggregate - OR - that the reader you are using to read SDE 9.3 data does not support fme_aggregate. Did you also WRITE to the SDE 9.3 in another workspace?

 

 

There is an option on the ArcSDE WRITER regarding Aggregates. Under ADVANCED options there is an option: Force in Aggregates" YES / NO. 

 

 

Which READER are you using on the 9.3 and 10 ?
Badge
Reader on the 9.3 is esri Geodatabase(ArcSDE Geodatabase) and reader on the 10.0 is Esri Geodatabase (File Geodatabase API). At this testing phase I am not using writing to some format, only input data to clip and output to Viever or Inspector. 
Badge
Anyone? Some tips?
Badge
What happens if you use ArcObjects to read the filegeodatabase (AKA the same reader for both databases) ?

 

The Filegeodatabase API reader is not using ArcObjects. Perhaps that is the difference?

 

/Tim

 

p.s. unless you are checking for fme_donuts and recieving fme_aggregates instead i doubt that there will be any problems. But both readers support both types.
Badge
I do realize that I mixed up few things at the same time. First, I thought, that new API fgdb reader has to be used If we use 10.0 or higher data version. That means NOT to use ArcObject. That was obviously a mistake. Second, I noticed that fme_geom for the same objects changed from donut (9.3) to aggregate (10.0) according to the different data version. Third, message of creating donuts happend only to 10.0 If data were tested at the one server we use for testing (installed FME2012,SP4), while testing it on the pc (FME2012 without SP) no such messages in log window). Fourth, the server we use for testing got some other problems at the same time which causes longer running time for FME. And I thought that this has to do something with data version or SP4.

 

And all this together, mixed up and shaked, made me confused. Thus, I have to use ArcObject as reader, update old workspace 2011 with new formatted transformers 2012, forget about wheter fme_geometry is donut or aggregate, and solve the other types of problems on testing server.
Badge +21
Also remember that you need to delete Transformers and Readers/Writers from workspaces generated with OLD FME, and add them again with the new Transformers Readers/Writers. If not, it will run with the "old" version of the Transformer/reader/writer. If you right click a Transformer and choose "Attach Summary" you see the version of a Transformer.
Badge +10
As per http://fmepedia.safe.com/articles/How_To/Transformer-Versioning
Badge
Sig Till, but If one has many workspaces, does it mean that this should be done everytime If new transformer appears? I assume what you will answer me:)  I think It should be some option/possibilty with new FME versions to update old tarnsformers with the message(warning) that this had been done or to manually choose it. 

 

Actually at this moment, I am not sure If I should remove old Clipper transf. with the new one, for instance. I don't need "Parallel Processing Level OptioN". Are there any other "invisible" changes within Clipper?!

Reply