I have the following code. I constructed an expression tree and I am stuck parsing it to find the result
You will find the details within my code
public enum OpertaionType { add, sub, div, mul}
public class Node {
public Node(Node lhs, Node rhs, OpertaionType opType, int index) {
this.lhs = lhs;
this.rhs = rhs;
this.opType = opType;
this.index = index;
}
Node lhs;
Node rhs;
OpertaionType opType;
int index;
}
class Program
{
static void Main(string[] args)
{
// I don't want this to be part of the node data structure
// because in the actual implementation I will end up referencing an
// array of data
int[] value = new int[5];
Node[] nodes = new Node[value.Length];
for (int i = 0; i < value.Length; i++)
{
value[i] = i+1;
nodes[i] = new Node(null, null, 0, i);
}
// suppose I constructed the tree like that
// note that the end nodes are marked by non-negative index that indexes the
// values array
Node a = new Node(nodes[0], nodes[1], OpertaionType.add, -1);
Node b = new Node(nodes[2], a, OpertaionType.mul, -1);
Node c = new Node(b, nodes[3], OpertaionType.div, -1);
// How can I find the result of Node c programatically
// such that the result is (a[2]*(a[0]+a[1]))/a[3] = 9/4
}
}
For a simple elementary intro to C# 3.0's own expression trees, see e.g. here; unfortunately I am not aware of a really broad and deep text on the subject (maybe a book..?).
As for your own hand-rolled format, you can evaluate it most simply by recursion; in pseudocode:
def valof(node):
if node.index >= 0:
return whateverarray[node.index]
L = valof(node.lhs)
R = valof(node.rhs)
if node.opType == add:
return L + R
if node.opType == mul:
return L * R
# etc etc
As a further twist, as you appear to want fraction results while the input values are integers, remember to use a fraction / rational number type for your computations - not sure if C# comes with one, but worst case you can find plenty on the net;-).
You need a recursive algorithm, passing the values array (code untested):
class Node{
//...
double compute(int[] values){
if(index >= 0)
return values[index]; // leaf node; value stored in array
switch(opType){
case add: return lhs.compute(values)+rhs.compute(values);
case sub: return lhs.compute(values)-rhs.compute(values);
case div: return lhs.compute(values)*rhs.compute(values);
case mul: return lhs.compute(values)/rhs.compute(values);
}
throw new Exception("unsupported operation type");
}
}
Notice that this performs all computations in double; if you really want 9/4, you would need to use a rational type.