delphi - String.Split works strange when last value is empty

Question

I'd like to split my string to array but it works bad when last "value" is empty. See my example please. Is it bug or feature? Is there any way how to use this function without workarounds?

var
  arr: TArray<string>;

  arr:='a;b;c'.Split([';']); //length of array = 3, it's OK
  arr:='a;b;c;'.Split([';']); //length of array = 3, but I expect 4
  arr:='a;b;;c'.Split([';']); //length of array = 4 since empty value is inside
  arr:=('a;b;c;'+' ').Split([';']); //length of array = 4 (primitive workaround with space)

question from:https://stackoverflow.com/questions/65937523/string-helpers-split-function-doesnt-consider-empty-spaces-after-the-last-deli

Answer 1

This behaviour can't be changed. There's no way for you to customise how this split function works. I suspect that you'll need to provide your own split implementation. Michael Erikkson helpfully points out in a comment that System.StrUtils.SplitString behaves in the manner that you desire.

The design seems to me to be poor. For instance

Length('a;'.Split([';'])) = 1

and yet

Length(';a'.Split([';'])) = 2

This asymmetry is a clear indication of poor design. It's astonishing that testing did not identify this.

The fact that the design is so clearly suspect means that it may be worth submitting a bug report. I'd expect it to be denied since any change would impact existing code. But you never know.

My recommendations:

1. Use your own split implementation that performs as you require.
2. Submit a bug report.

---

Whilst System.StrUtils.SplitString does what you want, its performance is not great. That very likely does not matter. In which case you should use it. However, if performance matters, then I offer this:

{$APPTYPE CONSOLE}

uses
  System.SysUtils, System.Diagnostics, System.StrUtils;

function MySplit(const s: string; Separator: char): TArray<string>;
var
  i, ItemIndex: Integer;
  len: Integer;
  SeparatorCount: Integer;
  Start: Integer;
begin
  len := Length(s);
  if len=0 then begin
    Result := nil;
    exit;
  end;

  SeparatorCount := 0;
  for i := 1 to len do begin
    if s[i]=Separator then begin
      inc(SeparatorCount);
    end;
  end;

  SetLength(Result, SeparatorCount+1);
  ItemIndex := 0;
  Start := 1;
  for i := 1 to len do begin
    if s[i]=Separator then begin
      Result[ItemIndex] := Copy(s, Start, i-Start);
      inc(ItemIndex);
      Start := i+1;
    end;
  end;
  Result[ItemIndex] := Copy(s, Start, len-Start+1);
end;

const
  InputString = 'asdkjhasd,we1324,wqweqw,qweqlkjh,asdqwe,qweqwe,asdasdqw';

var
  i: Integer;
  Stopwatch: TStopwatch;

const
  Count = 3000000;

begin
  Stopwatch := TStopwatch.StartNew;
  for i := 1 to Count do begin
    InputString.Split([',']);
  end;
  Writeln('string.Split: ', Stopwatch.ElapsedMilliseconds);

  Stopwatch := TStopwatch.StartNew;
  for i := 1 to Count do begin
    System.StrUtils.SplitString(InputString, ',');
  end;
  Writeln('StrUtils.SplitString: ', Stopwatch.ElapsedMilliseconds);

  Stopwatch := TStopwatch.StartNew;
  for i := 1 to Count do begin
    MySplit(InputString, ',');
  end;
  Writeln('MySplit: ', Stopwatch.ElapsedMilliseconds);
end.

The output of a 32 bit release build with XE7 on my E5530 is:

string.Split: 2798
StrUtils.SplitString: 7167
MySplit: 1428