Thanks for the explanation. However, I don't like how this works. I should simply be able to take any query that runs fine in other applications, and see/expect the same result when ran within JMP/JSL.
I've read though this thread and one thing is clear... the workaround discussed, splitting out all temp tables into separate queries, is NOT a good solution for us JSL Developers. I have a large, older query that works just fine and it will literally take me hours to split it out. What's the point of that? How is this the preferred method? I also noticed that in my first query split attempt, JMP isn't loading OIDs into the new Data Table. So I'm not really sure if I can split out this query. If I continue, I can't join on a blank OID column.... When I run this identical query portion through SQL Server, my column does have data.
I was also surprised that this topic was difficult to find -it took a few Google searches. Limitations should be clearly stated in the JSL Manual, but this is the first I'm reading it. I specifically wrote several helper functions -which now aren't very helpful. Now I'm realizing it was a waste of time. It will actually be faster for me to use another Application to write to csv, and have JMP load that. This software is not cheap, so I'd expect Temp Tables/SQL to work in JMP instead of needing an external app to save me time.
At the very least, I'd expect this application to have better notifications. If JMP detects temp table usage either within JSL or the "Open Table > Advanced" Platform, maybe a message should be sent to the developer to inform them that JMP cannot handle the operation...?
This thread doesn't appear active anymore... Does SAS/JMP plan on fixing this?