Skip to main content

I have created a workbench to detect schema different.

I am using two ChangeDector Transformer in this workbench which is producing output as I intend. Figure below shows that part.

Workbench for that part

 

This is output data (in csv file) based on that workbench.

Output by FME Workbench 2023.1 (FME Form 2023.1)


However, when I used the same workbench with the same data, the following output were produced. It got extra error. For error code E302, those not bold should not be there.

Output by FME Flow 2023.1.1.1

The workbench used is the same.
The data is the same.
Only output 1 run in FME Workbench 2023.1 (FME Form 2023.1) and output 2 Run on FME Flow 2023.1.1.1 and FME Flow 2023.2.

Note: The additional error contain field that is unchanged in first ChangeDetector and does not included in the second ChangeDetector (ChangeDetector3)

It could be related to different builds between your Form and Flow versions. Another thing you can try to rule out is if Flow isn’t processing multiple input files by mistake.

 


A couple of thoughts is that the results, given they are extra Features coming out of the “Insert” power tend to suggest there is actually a difference in the input data, with more Features read in the Flow version than the Workbench Version.

 

Looking at the output, it seems like there are not any unique Key values going into Original and Revised ports.  Whilst they are not mandatory to have, you can get some unintended behaviour from ChangeDetector if do not use a unique, non-Null, Feature Key Attribute to use as the Match/Compare Key on both the Revised and Original Features.  Looking at the data, should not “Field Name” or “Table Name”, “Field Name” be used as unique key values for the Features as it looks like comparing Schema to Schema?


Reply