The detailed category listing is dead simple to create: SELECT cat_id, cat_name, sum(amount) FROM category, line-item GROUP BY cat_name, cat_id;
But... as you say, the result is a LOT of numbers over the course of 160-ish categories of all levels. Which is why it is so important to be able to summarize. I want to summarize by the highest-levels of the category tree, precisely so as not to get a headache looking at the details. But it is still important to be able to drill down when I want more details.
On Sat, Feb 23, 2013 at 11:15 AM, Bryan L Nuse <nuse@uga.edu> wrote:
> That said, now that I have finally gotten the chance to try ltree, I think I like it a lot.
Hello Don,
Yes, after looking at ltree --which I had not done before-- I have to agree with Misa that it looks like the right solution for your problem. That is not to say that "brute force" SQL couldn't provide a workable arrangement; but ltree looks very flexible, especially as it allows you to assign cost values to non-terminal nodes. If it were me, though, I'd still make use of VIEWs to report results of the workhorse queries: staring at a list of items like "Transportation.Bicycle.Gear.Chain_ring" sounds like headache. That's a matter of taste, of course. Bryan
