Alan:
Thanks. This solution actually may work for me. Although I think the
recursive call in theory may run out of stack, my tree won't be close to
that kind of depth.
On Saturday, July 7, 2012 2:52:03 AM UTC-4, Alan Malloy wrote:
>
> If the value is something easy to compute, like a backreference in a tree,
> and you only need it while walking, why bother to store it at all?
> Especially if you're going to need to use mutable state, which will make
> your whole tree less useful for all kinds of things. Here's an example, in
> which I imagine that your walk wants to update each node so that its value
> is the sum of its old value, its parent's old value, and its children's old
> values - you can adjust what goes where if you want one or the other of
> these directions to see new values instead of old values.
>
> (defn transform [parent node]
> {:value (apply + (get parent :value 0) ;; the parent's value, or
> zero
> (:value node) ;; the node's old value
> (map :value (:children node))) ;; the children's sum
> :children (map (partial transform node)
> (:children node))})
>
> (transform nil {:value 1
> :children [{:value 2}
> {:value 3
> :children [{:value 4}]}]})
>
> {:value 6, ;; 1 for old value, 2+3 for children
> :children ({:value 3, ;; 2 for old value, 1 for parent
> :children ()}
> {:value 8, ;; 3 for old value, 1 for parent, 4 for children
> :children ({:value 7, ;; 4 for old value, 3 for parent
> :children ()})})}
>
> On Friday, July 6, 2012 4:56:22 PM UTC-7, Warren Lynn wrote:
>>
>> Hi, Phil:
>>
>> Thanks for your detailed analysis.
>>
>> My use-case is I have a tree and want to walk over the tree and apply a
>> function on the nodes. The operation of the function also depends on the
>> parent and children of the node, so it will be nice if each node already
>> contains the references to its parent and children. Something like that.
>>
>> I guess I will just use atoms to hold the reference. I am not in pursuit
>> of pure functional programming, so if it is too much trouble to do it with
>> FP, I will take whatever works.
>>
>>
>>
>>
>>
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to [email protected]
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
