C# String Replace safely

Question

do you have any ideas or tips how to replace a string safely?

Example:

string Example = "OK OR LOK";

Now i want to replace "OK" with true and "LOK" with false.

Example = Example.Replace("OK", "true");
Example = Example.Replace("LOK", "false");

Now the Result is: Example = "true or Ltrue";

The Result should be: Example ="true or false";

I understand the problem, but i have no idea how to fix this.

Thanks

@WiktorStribiżew: That's _not_ a duplicate. OP doesn't want to replace only whole words but also substrings. But better matching strings have higher precedence and should be replaced first. — Tim Schmelter, Dec 08 '16 at 11:11
@TimSchmelter: The order of alternatives does not matter with word boundaries. — Wiktor Stribiżew, Dec 08 '16 at 11:14
@WiktorStribiżew: OP hasn't mentioned that he wants to replace words, he's using `String.Replace` that also doesn't care about boundaries. The issue is that a replacement string could be a substring of another replacement string(like "OK" and "NOK"). Then the order of replacement matters. — Tim Schmelter, Dec 08 '16 at 11:17

score 6 · Accepted Answer · answered Dec 08 '16 at 09:51

You could replace the longest strings first, f.e. with this method:

public static string ReplaceSafe(string str, IEnumerable<KeyValuePair<string, string>>  replaceAllOfThis)
{
    foreach (var kv in replaceAllOfThis.OrderByDescending(kv => kv.Key.Length))
    {
        str = str.Replace(kv.Key, kv.Value);
    }
    return str;
}

Your example:

Example = ReplaceSafe(Example, new[] {new KeyValuePair<string, string>("OK", "true"), new KeyValuePair<string, string>("LOK", "false")});

excellent! this code fixed this problem too: string Example = "test OR test 2"; — Gerry Seel, Dec 08 '16 at 10:11

score 4 · Answer 2 · answered Dec 08 '16 at 09:50

4

Example = Example.Replace("LOK", "false").Replace("OK", "true");

answered Dec 08 '16 at 09:50

Rai Vu

1,595
1
20
30

score 3 · Answer 3 · answered Dec 08 '16 at 10:02

3

You could look for "LOK" first, but that won't help with the more general problem of not matching "POKE" and producing "PtrueE", and so on.

The following looks for word boundaries:

new Regex(@"\bLOK\b").Replace(
  new Regex(@"\bOK\b").Replace("OK OR LOK", "true"),
  "false")

A more flexible approach again is to both look for word boundaries and to identify which replacement one is doing in a match-evaluator:

new Regex(@"\bLOK|OK\b").Replace("OK OR LoK", m =>
{
  switch(m.Value)
  {
    case "OK":
      return "true";
    default:
      return "false";
  }
})

This is the approach least likely to run into further conflicts between different search keys.

answered Dec 08 '16 at 10:02

Jon Hanna

110,372
10
146
251

okay works fine but if i want to do this, nothing happend: string OK = "OK"; string replace = new Regex(@"\b" + OK + "\b").Replace(Example, "true"); – Gerry Seel Dec 08 '16 at 10:55
Your second `"\b"` is the code for the backspace character, it should be `@"\b"` or `"\\b"`. As well as that if you're constructing dynamcally you should use `@"\b" + Regex.Escape(OK) + @"\b"` so that way if your search string (`OK`) contains any characters that are special in regular expressions they're escaped to match the actual character. – Jon Hanna Dec 08 '16 at 11:03

Mong Zhu · Answer 4 · 2016-12-08T10:40:12.497

1

This might be an overhead for such a problem, but here is a version using RegEx with a negative lookbehind:

string Example = "OK OR LOK";
// Replace "OK" which is not preceded by any word character
string res = Regex.Replace(Example, @"(?<!\w)OK", "true");
string res2 = Regex.Replace(res, @"(?<!\w)LOK", "false");

Console.WriteLine(res);
Console.WriteLine(res2);

EDIT: inspired by @Jon Hanna.

if the OK or LOK should have a tail like OKE or LOCKS a positive looakhead for space (\s) or end of string ($) could solve the problem:

string res = Regex.Replace(Example, @"(?<!\w)OK(?=[\s|$])", "true");
string res2 = Regex.Replace(res, @"(?<!\w)LOK(?=[\s|$])", "false");

edited Dec 08 '16 at 10:40

answered Dec 08 '16 at 09:56

Mong Zhu

23,309
10
44
76

What happens with "OKE"? – Jon Hanna Dec 08 '16 at 10:03
1

`\b` is useful here. It matches the boundary between word characters and either non-word characters or the start and end of a string without the need for lookaheads. – Jon Hanna Dec 08 '16 at 10:52
@JonHanna AHH cool, good to know, thank you. Since you have it already in your answer, I won't duplicate it unnecessarily :) – Mong Zhu Dec 08 '16 at 10:58
Yeah, I just guessed you'd find it useful info. – Jon Hanna Dec 08 '16 at 11:04

C# String Replace safely

4 Answers4

Linked