Previously we were querying the full book objects just to get a list of
id's, which is much slower and also takes a lot more memory, which can
cause the process to be killed on memory-limited machines with a large
number of books.
Instead, since we're just dispatching jobs here, we can just ask for the
id's, which is faster and much more practical memory-wise.
The map is a little annoying, I didn't see a way to directly get just a
list of the value of one field, so we have to get a list of
dictionairies with one key and then pull that key out. Whatevs.