Arboreal is yet another extension to ActiveRecord to support tree-shaped data structures.
Arboreal surfaces relationships within the tree like
siblings as scopes, so that additional filtering/pagination can be performed.
It delegates as much work as possible to the underlying DBMS, making it efficient to:
fetch all ancestors, descendants or siblings of a node
move nodes (or subtrees) around
rebuild the hierarchy
First, install the “arboreal” gem, and add it to your Rails project's
Next, you'll need a migration to add
ancestry_string columns, and indices:
class MakeThingsArboreal < ActiveRecord::Migration def self.up add_column "things", "parent_id", :integer add_index "things", ["parent_id"] add_column "things", "ancestry_string", :string add_index "things", ["ancestry_string"] end def self.down remove_index "things", ["ancestry_string"] remove_column "things", "ancestry_string" remove_index "things", ["parent_id"] remove_column "things", "parent_id" end end
Finally, you can declare your model arboreal:
class Thing < ActiveRecord::Base acts_arboreal # .. etc etc ... end
Arboreal adds the basic relationships you'd expect:
In addition, it provides the following handy methods on each tree-node:
subtree(the node itself, plus descendants)
root(the topmost ancestor)
The first four return scopes, to which additional filtering, ordering or limits may be applied.
At the class-level:
rootsis a named-scope returning all the nodes without parents
rebuild_ancestryrebuilds the ancestry cache, as described below
Internally, Arboreal uses the
ancestry_string column to cache the path down the tree to each node (or more correctly, it's parent. This technique - a variant of “path enumeration” or “materialized paths” - allows efficient retrieval of both ancestors and descendants.
It's conceivable that the computed ancestry-string values may get out of whack, particularly if changes are made directly to the database. If you suspect corruption, you can restore sanity using
The ancestry rebuild is implemented in SQL to leverage the underlying DBMS, and so is pretty efficient, even on large trees.