Devin, is there a way to change the query to just pull all that do not have a link? In your example it is querying specific id numbers.

Sure, that's easy enough. The reason I provided a parent id, is so that you can find anything within a specific folder. Sometimes I'll search everything with a parent that has a specific parent so that I can search an entire folder tree.

SELECT tocid
FROM toc AS t0
LEFT OUTER JOIN toc AS t1 ON t1.tocid = parentid
WHERE (((
    SELECT COUNT(*)
    FROM toc AS t2
    WHERE t2.linkto = tocid
    )) = 0)    -- just remove the AND, 
               -- the zero is there because we're counting up the 
               -- documents that have a shortcut pointing to them,
               -- and we want to find the ones with none

Let me know if the SQL is too obtuse. I think better in C#, and tend to experiment and write my queries as LINQ statements in LINQPad and then clean up the generated SQL.

Devin, I understand now. We will need to actually exclude results that have specific parent ids. That gets a little trickier though as the main folder we want to exclude has 2 levels of subfolders.

This is another reason why it can be a good idea to use a staging table. For really complex cleanup jobs, I'll do a number of queries, and each time add the results to the temporary table. That way, my queries are simpler and easily validated. Then at the end, it's trivial to run a workflow or SDK script against the compiled data set.

Devin, is this the correct syntax?

SELECT t0.tocid
FROM toc AS t0
LEFT OUTER JOIN toc AS t1 ON t1.tocid = t0.parentid
WHERE (((
    SELECT COUNT(*)
    FROM toc AS t2
    WHERE t2.linkto = tocid
    )) = 0) AND (t1.parentid != 1)

It did not like just the tocid after SELECT or the parentid at the end of hte LEFT OUTER JOIN.

Yeah, you've got it. I messed up when I was reformatting the query.