Skip to main content
Hi All,

 

 

I'm working on a process that requires us to perform a comparision between an ArcSDE child version and its parent, determine what the differences are and extract the changes out (to a new version in a different database to be precise). From what I can tell, it looks like it would be possible to utilize the ChangeDetector transformer to do this, using the parent (SDE.Default) as the Original and the child as the Revised for all the feature classes. However, I am curious as to how well it would perform with larger feature classes (i.e. say a few feature classes with 1 million records each). Does anyone have any metrics for comparable tasks?

 

 

I'm guessing that the workbench performs a full compare between the two feature classes. Is there perhaps a better workbench to use in a versioned environment?

 

 

Thank you,

 

 

Chris
Hi,

 

 

the ChangeDetector, like most transformers in FME, is data source agnostic, meaning that the source of the input data usually doesn't influence the performance. On the other hand, the reader/writer and the underlying database can make a big difference.

 

 

I haven't worked much with ESRI versions in FME, but whenever I've had to use them I've noticed quite a performance penalty for some read/write operations. I suspect this is more a limitation of the ArcGIS libraries than FME.

 

 

David
Hi David,

 

 

Thanks for the response. You've warned me off performing the compare with FME :) Due to the versioning and the size of the datasets, I've decided against performing the actual compare within FME. There are some purpose built ArcObjects that can be used to perform the compare at the version level (and it looks like it uses the A and D tables directly for a faster response), which will then feed into my FME process post-compre.

 

 

Thanks again,

 

 

Chris

Reply