I'm an FME newbie and am trying to do the equivalent of a join in GIS, where I have a key field in my source data and a key field in my joining data. The issue here is that the fields in these layers are of different data types (source = int, join = char). I tried to use the featurejoiner for for this, but no joined attributes came across in the output. I'm assuming this is because they are of different data types. How can I get around this without having to write a new dataset that has the char attribute converted to int as an intermediary step? Is there a way around having to do this?
Solved
How to join features on attributes with different data types?
Best answer by carmijo
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
Right, its likely because the field names are the same and you have attribute conflict resolution set to use left for the featurejoiner. you can use a bulkattributerenamer before the psi join to add a prefix to the attributes coming from tester_25. This way they don't get overwritten or lost
--
If i've provided something helpful, please upvote the comment. If i've provided a solution, please mark it as 'Best Answer'. This helps the community find solutions when your question is referenced.
If I didn't provide a solution, please let me know if I can help further. Thanks! 🙂
+11- Supporter
Hey @lspencer, can you share the workspace, a sample of it, or at least show a screenshot of how you have the featurejoiner parameters configured? This would be helpful.
If it is the datatype difference that's causing the problem, you can use fme to temporarily adjust the datatype in the workspace before the features make it to the featurejoiner. The stringreplacer can help with this
carmijo wrote:
Hey @lspencer, can you share the workspace, a sample of it, or at least show a screenshot of how you have the featurejoiner parameters configured? This would be helpful.
If it is the datatype difference that's causing the problem, you can use fme to temporarily adjust the datatype in the workspace before the features make it to the featurejoiner. The stringreplacer can help with this
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
+11- Supporter
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
Thank you, can you attach a summary annotation to the feature joiner you've circled and share that? that will show the settings of the feature joiner
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
I'm pretty new to FME, do you mean you need a file from the SummaryReporter that is attached to the joined results? Or add in an annotation and describe what the featurejoiner settings are?
+11- Supporter
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
No problem. Welcome to the software and community, btw. If you right click the transformer, then choose show summary annotation, it will attach an annotation that summarizes the transformer parameters. this is an easy way to show others how the transformer is configured. 
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
Very cool and yes, super useful! So, I was doing the 1st join wrong...because you know, of course. The 2nd join is the one I'm having issues with now. Basically, I'm taking the Oracle table, splitting it into 2 parts, joining the 1st split to a GIS layer, then joining the 2nd split to the GIS using attributes in the already-joined Oracle table. SPECIFICATION_NO is the attribute for both sources for the 2nd join, where AssetTag (GIS) and ATTRIBUTE_VALUE (Oracle) are the attributes for the 1st join. 
+11- Supporter
- Best Answer
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
Right, its likely because the field names are the same and you have attribute conflict resolution set to use left for the featurejoiner. you can use a bulkattributerenamer before the psi join to add a prefix to the attributes coming from tester_25. This way they don't get overwritten or lost
--
If i've provided something helpful, please upvote the comment. If i've provided a solution, please mark it as 'Best Answer'. This helps the community find solutions when your question is referenced.
If I didn't provide a solution, please let me know if I can help further. Thanks! 🙂
lspencer wrote:
Hi @carmijo , here is my workspace - the values in the 2 tables are identical, but their datatypes are different. Any help appreciated, thank you!!!
Thank you for your help! Much appreciated, that did the trick 👍 Happy transforming!
Reply
Rich Text Editor, editor1
Editor toolbars
Press ALT 0 for help
Helpful Members This Week
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.