I had this issue and it drove me nuts. Completely blew away the association and even the objects referred to in the error message. Nothing after hours. After beginning it turns out the problem was occuring on a change activity that had nothing to do with the association in the stack trace.
Ultimately I was changing an object and setting the association TWICE in the same action
Hi Tracy,
It looks like this is happening in your local development HSQL database. I can't say how, but it looks like there's inconsistent data in the association called "taskline_request" in the "esat" module. Did you perhaps change this association from *-* to *-1 or 1-1 by chance?
One idea I have is to change this association to *-*, start the app, then change it back to whatever it should be. You might lose the data in this association (not 100% sure) but at least your other dev data will remain intact.
Another guaranteed solution is to wipe out your <project>/deployment directory and let Mendix rebuild it. That'll solve the error but you'll have to recreate any data you have in your development system.
Is the scenario that you use Java actions to commit large amounts of objects for performance reasons? In that scenario I had this before in some Mendix releases.
Mendix has fixed this issue in later Mx6 releases by adding some unique constraints on the id combinations columns of the association tables.