How to parse out Timestamp from this String?

I have this sample string:

Sample string 1:
A^1.1#B^1#I^2#f^0#p^1#d^2010-07-21T08:52:05.222ZKHBDGSLKHFBDSLKFGNIF#%$%^$#^$XLGCREWIGMEWCERG

Sample string 2:
A^1.1#B^1#f^0#p^1#d^2010-07-22T07:02:05.370ZREGHCOIMIYR$#^$#^$#^EWMGCOINNNNNNVVVRFGGYVJ667VTG

So, from these strings, I need to take out the time stamp:

2010-07-21T08:52:05.222 or
2010-07-22T07:02:05开发者_开发技巧.370

Basically values b/w d^ and Z

What is the best ("smartest") way to do this? substring(), regex?

Pattern p = Pattern.compile("(\\d{4}-\\d{2}-\\d{2}T\\d{2}:\\d{2}:\\d{2}\\.\\d{3})");
//you could also use "d\\^(.*)Z" as your regex patern
Matcher m = p.matcher("your string here");

if (m.find()) {
    System.out.println(m.group(1)); //print out the timestamp
}

Taken from here

Also, make sure to reuse the Pattern p object if you're looping through a series of strings

With two small assumptions you can do it without a regex.

1. The ^d right before the date string is the first one that appears in the text. I assume that delimiter always means "A date follows."
2. That date format looks pretty regular, so I'm assuming the length won't change.

Just get the index of the starting ^d delimiter to find out where the date starts and use the length to get then ending index.

public static void main(String[] args) {
    String s1 = "A^1.1#B^1#I^2#f^0#p^1#d^2010-07-21T08:52:05.222ZKHBDGSLKHFBDSLKFGNIF#%$%^$#^$XLGCREWIGMEWCERG";
    String s2 = "A^1.1#B^1#f^0#p^1#d^2010-07-22T07:02:05.370ZREGHCOIMIYR$#^$#^$#^EWMGCOINNNNNNVVVRFGGYVJ667VTG";

    System.out.println( parseDate(s1) );
    System.out.println( parseDate(s2) );
}

public static String parseDate(String s) {
    int start = s.indexOf("d^") + 2;
    int length = 23;

    String date = s.substring(start, start + length);
    return date;
}

Output:

2010-07-21T08:52:05.222
2010-07-22T07:02:05.370

I would go with a regex, something like (\d{4}-\d\d-\d\dT\d\d:\d\d:\d\d\.\d{3}).

You might want to get fancier, and prevent months outside the range 01-12, days outside of 01-31 etc. for hours but it should be good enough as is given the sample data that you've provided.

Substrings might work if the date is always prefixed with ^d, but I still think the regex is cleaner.