I have read many articles about dealing with hierarchies in postgres including nested sets, ltree, materialized paths, using arrays as parentage, CTEs etc but nobody talks about the following scenario.
Say I have a hierarchy like this
1
1.1
1.1.1
1.1.2
1.2
1.3
2
2.1
In this hierarchy the order is very important and I want to run frequent(ish) re-ordering of both subsets and entire trees and even more frequent inserts.
Scenario 1: I want to insert a child into the 1.1 subtree. The next item should be 1.1.3 and I can't figure out any other way to do this other than to subquery the children and to figure out the max child ID, add one to it which is a race condition waiting to happen.
Scenario 2: I now decide the recently inserted item is the second most important so I reset the ID to 1.1.2 and then increment 1.1.2 (and possibly everything below). Again this is both prone to race conditions and involves a heavy update.
Is there a better way to deal with this or is the complexity unavoidable?
I should state that like most database reads will be much more frequent than writes and inserts will be more frequent than updates (re-ordering)