Maybe we still need to find better terms than »database node« and »fragments« to describe the data structures that we use for XML structures (as described in [1]), as the current situation is even more complex:
1. we have fragments (e.g. as the result of <x/>) 2. we have volatile main-memory database nodes 3. we have persistent database nodes
…and db:node-id can only be used for persistent database nodes.
In my case, I’m using transform with to augment content pulled from the database before doing processing that will vary depending on whether I have a database node (i.e., getting its node ID or database path) or I don’t.
Feel free to provide us with a little use case.
[1] https://docs.basex.org/wiki/Database_Module#Database_Nodes