What is wrong with my coding?

Question

This is a fragment of a class that removes repeated values in fArr.

procedure tArray.removedup; 
var
  K,R : word;
  DArray : array of string;
begin
  SetLength(DArray, max);

  for K := 1 to max do
    if fArr[K] <> '' then
    begin
      fCount := fCount + 1;
      DArray[K - 1] := fArr[K];
    end;//if

  SetLength(DArray, fCount);
  K := 1;
  while K < fCount do
  begin
    for R := K+1 to fCount do
      if DArray[K] = DArray[R] then
      begin
        finalize(DArray[R]);
        if fCount - R > 0 then
          Move(fArr[R + 1], DArray[R], SizeOf(string) * (fCount - R));
        Initialize(DArray[fCount - 1]);
        SetLength(DArray, fCount);
      end; //if

    K := K + 1;
  end;//while

  for K := 1 to fCount do
    fArr[K] := DArray[K-1];
end;

I know it looks messy... but the problem is actually at the last for loop. The program bombs out saying "Invalid pointer operation"

Answer 1

There are many things wrong with your code. It's difficult to know where to start. There is excessive copying. Much confusion over zero-based array indices. And worst of all, that call to Move that goes behind the compiler's back and copies managed objects without the compiler getting a chance to reference count properly.

Any of these problems could be responsible for your runtime error. I won't attempt to debug or fix your code. I don't think anything useful can be salvaged from it. Here's a function that removes duplicates:

function Deduped(const input: array of string): TStringArray;
var
  i, j, Count: Integer;
  Found: Boolean;
begin
  Count := 0;
  SetLength(Result, Length(input));
  for i := 0 to high(input) do begin
    if input[i] <> '' then begin
      Found := False;
      for j := 0 to Count-1 do begin
        if input[i] = Result[j] then begin
          Found := True;
          break;
        end;
        if not Found then begin
          Result[Count] := input[i];
          inc(Count);
        end;
      end;
    end;
  end;
  SetLength(Result, Count);
end;

I'm ignoring empty strings as I think your code attempts to do.

The function is simple enough. It allocates an array large enough to hold all the values. Then it works through the input values and for each one checks if it has already been stored in the output array. If not then it adds the value to the array. Finally the output array is resized to the final length. That is the number of unique non-empty strings that were found.