A naive approach would be to split your URL into groups (say url.split("/")) and compare the resulting arrays. It the parts match, just add them as constant strings to the pattern. If they don't, add a pattern that matches all possible values. Here is a simple implementation:
public static void main(String[] args) throws Exception {
String[] urls = {
"http://www.buy.com/prod/disney-s-star-struck/q/loc/109/213724402.html",
"http://www.buy.com/prod/samsung-f2380-23-widescreen-1080p-lcd-monitor-150-000-1-dc-8ms-1920-x/q/loc/101/211249863.html",
"http://www.buy.com/prod/panasonic-nnh765wf-microwave-oven-countertop-1-6-ft-1250w-panasonic/q/loc/66357/202045865.html",
"http://www.buy.com/prod/escape-by-calvin-klein-for-women-3-4-oz-edp-spray/q/loc/66740/211210860.html",
"http://www.buy.com/prod/v-touch-8gb-mp3-mp4-2-8-touch-screen-2mp-camera-expandable-minisd-w/q/loc/111/211402014.html"
};
String all = "[^/]+";
String[] pattern = urls[0].split("/");
for (int i = 0; i < urls.length; i++) {
String parts[] = urls[i].split("/");
// TODO handle urls with different number of parts
for (int j = 0; j < pattern.length; j++) {
// intentionally match by reference
if (pattern[j] != all && !pattern[j].equals(parts[j])) {
pattern[j] = all;
}
}
}
// build pattern - use [^/]+ as a replacement (anything but a '/')
StringBuilder buf = new StringBuilder();
for (int i = 0; i < pattern.length; i++) {
buf.append(pattern[i] == all ? all : Pattern.quote(pattern[i]));
buf.append("/");
}
// stip last "/"
buf.setLength(buf.length() - 1);
// compile pattern
Pattern p = Pattern.compile(buf.toString());
// output
System.out.println(p.pattern());
for (int i = 0; i < urls.length; i++) {
System.out.println(p.matcher(urls[i]).matches());
}
}
Here's the output of this example:
\Qhttp:\E/\Q\E/\Qwww.buy.com\E/\Qprod\E/[^/]+/\Qq\E/\Qloc\E/[^/]+/[^/]+
true
true
true
true
true
As you see, the pattern looks a bit weird. That's due to the Pattern quoting. Nevertheless, the pattern matches all urls from this example. There's some work left though, most noteably handling urls with different number of parts after split and common suffixes (.html).