Question

Can any body give me some regular expression patterns which will extract the version numbers(1.5) from some sample strings like

    "jdk1.5","jdk-v1.5","jdk-V1.5","jdk V1.5","jdkv1.5","jdk version1.5","1.5.6","v1.5","V1.5","version 1.5","Version 1.5","14.5.4","1.5.4","14.5.4""jdj14.5"

I want to store the regex patterns in a string array and will check these patterns with the above strings. If there is a match with any of the stored patterns, then the output should be the version number 1.5. I want to extract the version numbers (1.5) only from the above strings.

valid string formats :
“jdk1.5″,”jdk-v1.5″,”jdk-V1.5″,”jdk V1.3″,”jdkv1.4″,”jdk version1.5″,”1.5.6″,”v1.5″,”V1.5″,”version 1.5″,”Version 1.5″,”14.5.4″,”1.5.4″,”14.5.4″”jdj14.5″,”14.52.3.42”

Invalid String formats : “jdk1..2″,”jdk.1.2.”,”.1.2.”

Answer 1

Requirements

If you can formulize what you want as:

1. Anything in format 1.5 or 1.5.6 or 1.5.6.7 or 1.5.6.7.8 etc. is valid version number by default (numbers can be one or more digits like 12.60.70).
2. A version number 1.5 can be followed by a dot (.) only if it is in the format 1.5.6.
3. A version number 1.5 never preceded by a dot (.) like .1.5.

Solution

Then, I can suggest you this regex:

(?!\.)(\d+(\.\d+)+)(?![\d\.])

See it in action, includes all positive samples and excludes all negative samples you provided.

How to use

First capture group will be your version number. Sample code:

Pattern pattern    = Pattern.compile("(?!\\.)(\\d+(\\.\\d+)+)(?![\\d\\.])");
Matcher matcher    = pattern.matcher(inputStr);
boolean matchFound = matcher.find();

if (matchFound)
{
    String version = matcher.group(1);
    System.out.println("Version number: " + version);
}
else
{
    System.out.println("No match for the input!");
}

Note: This will work only with Java 7+, because look-aheads doesn’t supported by older versions.