DataSet.Find() is abit dodgy

[da288]

...or perhaps it's just my AS! Anyway I'm stuck again and turning to you guys for help.

After lots of debugging I found that dataset.find() only lets you search a dataset with 1 value accurately (specifying multiple values which it is obvious exist in the dataset keeps returning "not found").

However my main problem that I'm posting about here is the complete opposite: the dataset is returning "found" even when the value doesn't exist!

Here is my AS:

if {//dataset length is 0, add a value to the dataset }
else { // so the dataset length is not 0
      trace("finding context: "+contextvalue);
        if(contextlist3.find([contextvalue])) { // if found
      trace("found context");
        existing=contextlist3.id;
      trace("id found: "+existing);
      return existing;
        }

        else { //if not found
       trace("getting to newpos");
}

I realize this is very vague but I think it just may be my if/else structure? I even tried a "if not":

if(!contextlist3.find([contextvalue])) {
trace("not found");
}
else { // do something else }

but it still branches to the "else".

I realize there is an adobe notice regarding binding and the dataset bugs but mine is bound correctly, I can add values etc fine but it always returns a "found"... really wierd. Any feedback/advice much appreciated!

Update: I have discovered that if I try searching for a string it exhibits the above behaviour, but if it's a numerical value it doesn't. I've been reading about the datatype.setAsString or something but I have no clue what it's on about. Any ideas?

The values I am searching for are:
"00000000-0000-0000-0000-000000000001"
and "Unknown"

and they both are apparently equal according to dataset.find.

[da288]

Ok after hours of debugging found the solution:

1)You can only accurately search a field that you have sorted using ADDSORT.
2)(Not so sure about this one but from what it looks like) - you must specify all Sorted fields, in order, in the search term. e.g. to search a dataset with the following structure with 2 columns titled Field1 and 2 populated with Value 1 and 2:

Field1  | Field 2
Value1  Value2

addsort must have both Field1 and Field2 specified.
and to search the dataset: dataset.find([Value1, Value2]) must both be specified (could be wrong on this one - hope I am!).

Hope that helps for anyone stumbling accross this in teh future.