OpenEuphoria

Re: limitations of sequences, revisited

• Posted by David Cuny <dcuny at LANSET.COM> Dec 27, 2000
• 596 views

George Henry wrote:

> No! And my "complaint" was that I cannot
> use such techniques to MODIFY a value deeply
> buried within the structure.

I think the confusion stems from the fact that, in Euphoria, it's often
trivial to embed all the data in the same structure. And it's often
convenient to do just that.

But for cases where this is problematic (like your example), you can
trivially model the code as it is done in C by using indexes instead of
pointers.

For example, you could create the tree as an array of tuples:

constant
    DATA   = 1,
    LEFT   = 2,
    RIGHT  = 3

Instead of storing a subtree in the left and right nodes, store the index to
the node. Here's a trivial example showing a binary tree for the data values
"a", "b", "c" and "d":

{
   { "b", 2, 3 },
   { "a", 0, 0 },
   { "c", 0, 4 },
   { "d", 0, 0 }
}

Each node of the tree can be identified with a single index.

Or you could choose to seperate the data from the structure. The following
example is the same as the prior, but instead of placing the data in the
structure, indexes to the data (which is placed in another structure) are
used:

-- structure
{  { 1, 2, 3 },
   { 2, 0, 0 },
   { 3, 0, 4 },
   { 4, 0, 0 }
}

-- data
{ "b", "a", "c", "d" }

Another option is to keep the embedded tree structure, but store the data in
a seperate structure. The following example embeds a tree in each node, but
stores the data in a seperate structure. Instead of storing the data in the
tree, the first element of the tuple contains an index to the data:

   { index to data, left node, right node }


-- the tree
{1, {2, {}, {}}, {3, {}, {4, {}, {}}} }

-- the data
{ "b", "a", "c", "d" }

In any event, by either flattening the structure and using indexes, or
storing data in a second sequence and referencing them with indexes, you can
accomplish what you want without having to resort to pointers.

I hope this helps!

-- David Cuny