I have to agree strongly that a table is a far better container for items between job steps when not using SSIS, and for the following reasons:

1.) It's easy to see within the job steps how different tasks get data from or to each other.

2.) It's a lot easier to see what happened if intermediate results are in a table when a job fails in the middle.

3.) It's much easier to document and explain, and someone coming in cold could probably quite easily figure it out without having to go to BOL or do research.

4.) With the right design, you might even be able to just rerun the entire job when it breaks, and it will figure out it's state from such a table.

Great Idea !!!

Steve

(aka smunson)

:-):-):-)