String.replaceAll does not apply to any kind of characters

Question

So what i've got is this simple method which should remove all characters except upper and lowercase letters and numbers from a string.

public static String toSimple(String arg) //redurziert einen String auf Buchstaben und ganze Zahlen
{
    String string = arg;
    int i = 0;
    while ( i < string.length())
    {
        if (((int)string.charAt(i) >= 48 && (int)string.charAt(i) <= 57)||((int)string.charAt(i) >= 65 && (int)string.charAt(i) <= 90)||((int)string.charAt(i) >= 97 && (int)string.charAt(i) <= 121))
            i+=1;
        else
        {
            int a = string.length();
            string = string.replaceAll(""+string.charAt(i), "");

            if (!(string.length() < a)) //Just in case
            {
                i+=1;
            }
        }

The problem is that in some cases string.replaceAll won't change anything though reached and given a character. I checked it in the debugger and couldn't find any apparent errors or exceptions. I didn't check the whole ASCII table, I know that the problem will occur for '$' '?' and '.'. Any suggestions how to solve that?

Note that char literals can be used instead of "magic numbers" - so `(int)string.charAt(i) >= 48` can be written as `string.charAt(i) >= '0'`, which makes it a lot more obvious what you mean. (The cast to `int` isn't necessary in any case). — Andy Turner, Nov 08 '16 at 08:03
That's a pretty horrible way to remove characters from a String. And it doesn't work. Try the string ",,,+a" - it won't remove the "+". Normally you would build a new String in a StringBuilder but even if you were doing it like this, you'd use `String.substring` twice to get the string before and after the character. — Erwin Bolwidt, Nov 08 '16 at 08:05

Andy Turner · Answer 1 · 2016-11-08T08:05:45.167

2

String.replaceAll takes a regular expression as its first parameter.

$, ? and . are special characters in regular expressions.

Use String.replace instead - this actually uses regular expressions internally too, but it escapes the strings correctly.

However, given that you're using String.replaceAll, you can just use a regular expression to do the whole thing:

return arg.replaceAll("[^A-Za-z0-9]", "");

edited Nov 08 '16 at 08:05

answered Nov 08 '16 at 08:01

Andy Turner

137,514
11
162
243

man, that's much easier ':D thanks – Felix Stiep Nov 08 '16 at 08:05
1

Or `arg.replaceAll("\\P{Alnum}+", "")` – Wiktor Stribiżew Nov 08 '16 at 08:09
@Wiktor indeed. I suggested the explicit ranges because that matches OP's code. – Andy Turner Nov 08 '16 at 08:10

String.replaceAll does not apply to any kind of characters

1 Answers1