Hi,
I'm using a FeatureMerger to check that two files contain the same features based on their attributes.
Apparently, they don't seem to contain the same features although I thought they would. I must be missing something in the data. That's allright, FME's there to help me isn't it :)
But strangely, the features which are not merged are not output by the UnmergedRequestor port of FeatureMerger. Isn't that weird ?
Hey. I solved the problem with a NullAttributeMapper transformer. But still, I don't understand why the features weren't output through the UnmergedRequestor port.
@timotheebecker If you can attach the sample data we can look into the why's in more detail.
Hi @markatsafe.
Here attached is a workbench and data. I have come against the same problem again with different datasets.
Thanks in advance.
@timotheebecker Thanks for the reproduction data & workspace. I couldn't reproduce the issue in FME 2021.2:
Perhaps try changing the Excel reader behaviour to treat empty cells as <null> rather then the default <missing> . <nulls> are generally treated as real citizens in FME were as <missing> implies the attribute doesn't exist.
If that doesn't work - can you identify the 5 records that do not join?
Hi @Mark Stoakes !
Yes, what you suggest works ! I suppose it's even better than what I had done : adding NullAttributeMappers (which do the exact same thing).
I suppose that is a good practise with Excel files to choose the option "Read blank cells as Null".
But I still find it weird.
Here below is a screenshot of an even simpler testcase. I attempt to merge two times the same file (just duplicated and named differently).
And it still doesn't merge when Read blank cells is set to Missing. I suppose this is "As designed". I wonder what the logic behind is...
@Mark Stoakes suggeted to use the avanced reader parameter "Read blank cells as :" null instead of missing. And that solves the problem. strangely, I don't see his post anymore...