is matches() slower than the char array check in String validation?

[Robin John]

is this slower ???

  
if (myString.matches("[a-zA-z0-9]*")) {
    System.out.println("Alphanumeric");
} else {
    System.out.println("Not alphanumeric");
}

than this ???

  
//pseudocode
char[] c = myString.toCharArray();

for(each charC in c){
  if(charC < '0' || charC > '9')
    break and declare not numeric
} 
declare numeric

[steve souza]

1) I suspect this would never be a bottleneck in a real application so it makes more sense to do which makes your vailidation the most flexible and easiest to maintain. Either way I would come up with a validate(...) method and hide the implementation.
2) If you really want to see performance differences run a lot of validations and time each approach.
3) You can also look at the source code. Here is a good site that has the java source. String.matches calls Pattern.matches. http://www.java2s.com/Open-Source/Java-Document/6.0-JDK-Core/Collections-Jar-Zip-Logging-regex/java/util/regex/Pattern.java.htm

[Pat Farrell]

Robin John wrote:

  
//pseudocode
char[] c = myString.toCharArray();

for(each charC in c){
  if(charC < '0' || charC > '9')
    break and declare not numeric
} 
declare numeric

Does your code work internationally? Its probably fine for ASCII, but the A in ASCII is American.

A major point of the generalized pattern stuff is to fit well into international code, I tend to think that languages such as Arabic or Tamil are not going to use American digits.