I have a datastage job that is comparring a previous dataset with a current one and loads the values into another dataset. The problem is that it does not search for the correct previous dataset and fails to see the physical file. This happens to only a part of the files and is not a parameter problem as all the parameters are project level and set up correctly.
Is there a possibility of being overwriting? What type of error do you get? Was the dataset correctly created? Is the file system asociated full? There are many possibilities, that could be happening to you.