String.Split with newlines

General
1

I need to split a string based on UNIX newlines. What I’m trying to do is end up with a String array of the actual lines of text.

Var s1 As String = EndOfLine.UNIX
Var s2 As String = EndOfLine.UNIX + EndOfLine.UNIX
Var s3 As String = "Hi" + EndOfLine.UNIX + "you"

Var a1() As String = s1.Split(EndOfLine.UNIX) // Two empty elements
Var a2() As String = s2.Split(EndOfLine.UNIX) // Three empty elements
Var a3() As String = s3.Split(EndOfLine.UNIX) // Two elements containing "Hi" and "you"

Break

This to me doesn’t seem like the correct behaviour. What I would expect to see is:

// a1 = [""]

// a2 = ["", ""]

// a3 = ["Hi", "you"]
2

The function splits ON the character and the character is lost.

Your comment descriptions next to the lines are correct, the expectations for the empty elements are off by one.

// One EOL to split on, two elements
// a1 = ["", ""]

// Two EOLs to split on, three elements
// a2 = ["", "", ""]

// One EOL to split on, two elements
// a3 = ["Hi", "you"]
1 Like
3

What I expect is:

  • with no endofline.unix - 1 element
  • with one endofline.unix - 2 elements
  • with two endofline.unix - 3 elements
    no matter if the elements are empty or not
1 Like
4

Hmm. Does anyone have a good suggestion then to achieve what I’m trying to do? If it splits on the character and then loses it I don’t think that’s the correct approach to achieve what I’m hoping. It’s not like I can simply count the size of the resultant array and subtract one and that will give me the correct number of lines (because I don’t think it does). For example, a1 and a3 both have two elements but s1 is only one line and s3 is two lines.

5

you can simply String.TrimRight(EndOfLine.UNIX) and then do the split ?

Var s1 As String = EndOfLine.UNIX
Var s2 As String = EndOfLine.UNIX + EndOfLine.UNIX
Var s3 As String = "Hi" + EndOfLine.UNIX + "you"

s1 = s1.TrimRight(EndOfLine.UNIX)
Var a1() As String = s1.Split(EndOfLine.UNIX) // Two empty elements
System.DebugLog "s1 = " + String.FromArray(a1, "-")

s2 = s2.TrimRight(EndOfLine.UNIX)
Var a2() As String = s2.Split(EndOfLine.UNIX) // Three empty elements
System.DebugLog "s2 = " + String.FromArray(a2, "-")

s3 = s3.TrimRight(EndOfLine.UNIX)
Var a3() As String = s3.Split(EndOfLine.UNIX) // Two elements containing "Hi" and "you"
System.DebugLog "s3 = " + String.FromArray(a3, "-")

Break

DebugLog (macos):

s1 =
s2 =
s3 = Hi-you

Build a function that does Trim + Split.

6

Does your editor offer text wrapping? I would think this method of splitting by endofline wouldn’t be sufficient.

7

That’s a can of worms.

My current editor (XUICodeEditor) doesn’t but I do want to support that in the new one I’m working (at least eventually).

I’m still in the prototyping phase - trying to figure out the best way to store the text and line starts/finishes.

I’m currently leaning towards storing all the text in a gap buffer (line endings as well) and probably storing the start and finish offset of each line in a TextLine class (which are stored in an array in a LineManager class of the editor). By storing all the text in a gap buffer in theory that gives me the chance to soft wrap if needed.

8

As I understand it

Splitting on a character “that is included” in the string implies that there are two parts, before AND after.

Var s1 As String = EndOfLine.UNIX (thus has two parts)
“” before and “” after

if it was
Var s1 As String = “”
then you would have one part, which would be “”

Your s3 only has two parts because you only have 1 actual endofline.unix

Your s2 has 3 parts because there are 2 endoflines
the first part of s2 splits it into
“”, endofline
that remaining endofline splits into
“”, “”
The result is
“”, “”, “”

When doing a split you will always end up with n + 1 the number of times the character you are splitting on ACTUALLY is in the string. You can’t imply more.

1 Like
9

Another way to think of it would be like electric poles and wires. Electric poles would be the character(s) to split and the wires are the returned parts.

1 Like
10

Thanks for the explanations guys - they’re really helpful, particularly the wires and poles analogy.

I’m going to create a new topic about managing lines since it’s slightly different that this one.

11

In fact, it is a matter of “Number of Delimiters” and “Number of Intervals”.

Think:
I was born on a Mardi Gras.
On some rare years, Mardi Gras is also my Birthday. On these years, I am n years old, and n+1 Mardi Gras.

At first it does not seems obvious…