Matcher.matches() Vs. Matcher.find()

[spudillon]

hey folks,

As the title suggests this is a performance issue i have with the Machter object. I am writing a simple script to pick the times of particular events in a log file. In my log processor class i have a method isTimeStamp() that performes a simple check to ensure i have picked up a timestamp and not some unfortunate list of digits.

String regexTS = "\\d{1,2}[.:]\\d{1,2}[.:]\\d{1,2}[.:,]\\d{3,}";

public static boolean isTimeStamp(String arg) throws NumberFormatException {

boolean timestamp = false;
Pattern pattern = Pattern.compile(regexTS);
Matcher matcher = pattern.matcher(arg);

if(matcher.matches()) { // or i can use if (matcher.find()){
int i = matcher.start();
String[] tmpArr = arg.substring(i).split("[,.:]");
try{
if (Integer.parseInt(tmpArr[0])<24 &
Integer.parseInt(tmpArr[1])<60 & Integer.parseInt(tmpArr[2])<60){
timestamp = true;
}
}catch (NumberFormatException e){
System.out.println("EOD_Report: NumberFormatException thrown:"
+" LogProcessor.isTimeStamp(): arg is "+arg);
}


In tests between the two options i have found the .matches() method to be more efficient can anyone please confirm this and if possible explain this please.

Thanks

Spudillon

[eRaaaa]

both are different match operations so you cant compare them in my opinion.
Read Matcher (Java Platform SE 6)

But in your code, you don`t need the Matcher and Pattern class. You could use the matches method of the String class (String use Pattern and Matcher intern too...)

		if (arg.matches(regexTS)) {
			String[] tmpArr = arg.split("[,.:]");

Or use a better/other regular expression then you could avoid the rest of your code (parseInt , compare and so on) :D

[spudillon]

Thanks for taking the time to check out my query.

->"both are different match operations so you cant compare them"
<-point taken there is no generic answer so i have opted for the code "as is above".

->"But in your code, you don`t need the Matcher and Pattern class"
<-compared to String.matches() this is costly but necessary as i need matcher.start()
as i need to read multiple logs and timestamps are not always hh:mm:ss.mls(nsnsnsns)
some are prefixed with date YYYYMMDD-hh:mm:ss.mls and need to enure hh<24,
mm<60 and so on.

***If it helps anyone in the future in this case i found Matcher.matches() was more efficient than Macther.find() and this result is completely dependant on the example.