Skip to main content

I have two tables for blocks
data, each block in these tables is identified by its block ID, and I want to
find the unmatched blocks between these tables.

In ArcMap, I joined the two tables based on the "Block
ID" fields, after validating the join, 822 out of 1496 blocks were
matched.

However, I tried to do the same
thing using FME, I created a workbench that reads from these two tables, and
applied a feature merger on them to compare the "Block ID". However,
the results were different; there was 1011 matched "merged" blocks!

I could not figure out why FME
results are different from ArcGIS ones. What might be the issue here?

Best Regards,

Hani

Could this because of a difference in length of the BLOCK_ID field in both tables. And then FME trimming spaces from the values?


It could be a number of things. I think the important part is to try and identify one of the features that FME thinks is a join but ArcGIS does not. If you can inspect one of those features (and the feature it allegedly matches to) it will help to clarify what the issue is.


I agree with Mark, try to find one single feature that gets a different treatment in FME and ArcGIS.

That said, I kinda suspect an encoding issue here, seeing as you're matching on (what looks like) arabic characters.


Hi

Like Mark2 hinted, unicode data might be compared by collation order in a geodatabase but byte order by FME, which would cause match differences.

I'm going to pollute the forum with a pointer to an Esri sample that should help you, but hey I do plenty for Safe in my day job :-)

http://www.arcgis.com/home/item.html?id=e638afe069...


Not sure if this ever got resolved. @hani_draidi -- could you send some sample data and the workspace in to support@safe.com and we'll dig in.


Reply