是否有任何有效的Java解决方案/算法可以根据带分隔符的键字符串在文本字段中查找值？

Question

Is there any efficient Java solution/Algorithm to find values in a text field based on key string with delimiter?

Example:

Rules:

I. Key: username Start-Delimiter: ; End-Delimiter: ;

Need to find a value which if followed by key( in this case username ) and value is between start and end delimiters(in this case ;).

II. Key: on Start-Delimiter: ; End-Delimiter: ;

Sample Input:

A user with username ;suren; logged into the system on ;Thu May 2, 2013 2:30pm;

Results: I. username - suren II. on - Thu May 2, 2013 2:30pm

Here is an example of what I have tried so far:

String key = "username ";
String startDelimiter = ";";
String endDelimiter = ";";
String computedPattern = "(\\w+)" + key + startDelimiter + "(\\w+)" + endDelimiter;
Pattern p = Pattern.compile(computedPattern);
Matcher m = p.matcher(
  "A user with username ;suren; logged into the system on ;Thu May 2, 2013 2:30pm;");
while( m.find() ) {
  System.out.println( "Key:" + m.group(1) + "Value:" + m.group(2) );
}

Answer 1

The problem in your approach arises from your construction of the regex. If you look at Java Docs for the Pattern Class , you see that \\w matches just word characters: [a-zA-Z_0-9]

It sounds like you just want to match any characters between semicolons after a defined key with any number on spaces in between. The right pattern for this is the following:

ArrayList<String> keys = new ArrayList<String>();
keys.add("username");
keys.add("on");
String startDelimiter = ";";
String endDelimiter = ";";
String searchStr = "A user with username ;suren; logged into the system on ;Thu May 2, 2013 2:30pm;";       
for (String key : keys) {
    Pattern p = Pattern.compile("("+key+")[ ]+?"+startDelimiter+"([^;]+)"+endDelimiter);
    Matcher m = p.matcher(searchStr);   
    while( m.find() ) {
        System.out.println("Key: "+m.group(1)+" Value: "+m.group(2));
    }
}

The character class [^;]+ does the job for you, since it matches any non empty character sequence with characters different from ";" .