Java regex for html select tag parsing

Question

I will be writing a program where I just read a jsp file line by line and using regex check if there are any select tags in the page. If they are then I want to extract the data between <option>DATA</option> tags.

I wrote the following regex here at regex101.

(?:<select.*>\n?\s+(.*<option.*>$.*)\n?\s+<\/select>.*)

Those who are not aware about regex101, you can also see the java code for the regex using sidebar at top left.

Problem is regex is not matching any thing for the string:

<select name="javaType">
  <option value="boolean">boolean</option> 
  <option value="byte">byte</option> 
  <option value="char" selected>char</option> 
  <option value="double">double</option> 
  <option value="float">float</option> 
  <option value="int">int</option> 
  <option value="long">long</option> 
</select>

I want to extract the data between tags. I know my regex is incomplete but I have no clue on how I should proceed on this.

Do I parse line by line and check for select tag first, then again check for tags? Or is there a better way using regex?

I can't for some reason, I only have the option of using regex. — cFrags, Aug 11 '18 at 08:17
"I can't for some reason" and what reason is that? Are you getting some problems while using it (errors/exceptions/incorrect results) or is it requirement from someone else (who probably didn't see https://stackoverflow.com/a/1732454/1393766, or [Using regular expressions to parse HTML: why not?](https://stackoverflow.com/q/590747) or [Can you provide some examples of why it is hard to parse XML and HTML with a regex?](https://stackoverflow.com/q/701166))? Can you show us your attempts using parser (maybe you missed some mistake on your part which we somehow will find)? — Pshemo, Aug 11 '18 at 08:30
It's because of our client who is insisted on using regex only for this, I don't know what his problem is, but this is what was communicated to me — cFrags, Aug 11 '18 at 09:01
From my experience mass downvotes are often caused by belief that people shouldn't support choosing incorrect tools for the job (just like using `null` as layout manager in swing). HTML isn't regular language so it shouldn't be handled with *regular* expressions (see my earlier links). Proper tool here would be HTML parser. The more you use incorrect tool, the bigger your problem will be when you will face situation which will require from you moving to correct tool (and it almost certainly will happen since "the only thing that is constant is change"). So why not use proper tool at start? — Pshemo, Aug 11 '18 at 10:04

Dongfang Qu · Answer 1 · 2018-08-13T01:20:34.947

0

why don't you use some html parser to process it? your question is not very specific though.

I thought you might want to check which option has been selected. I give a simple regex base on my understanding.

selected\>(?<selected>(\w+))<

The named pattern of selected is what has been selected.

edited Aug 13 '18 at 01:20

answered Aug 11 '18 at 08:22

Dongfang Qu

341
1
11

For some reasons, its not possible for me to use html parser's, I have only option of using regex or something that won't require any additional jars in my project. – cFrags Aug 11 '18 at 08:59

score 0 · Answer 2 · answered Aug 11 '18 at 08:41

You are using look aheads(?:) in your regex to find the pattern. For this to work, all content needs to be in a single line. If you wanted to find a regex for bash then we could have used tr -s '\n' ' ' to transform the string in one line and then maybe look ahead would have helped.

We can do something similar in java also but I would suggest you to loop through file lines one-by-one, search for <select and once you find it, have a regex to find content between option till you find `.

Hope below code helps

public static void main(String[] args) throws IOException {
    String strTest = "D:\\Temp.jsp";
    File file = new File(strTest);
    BufferedReader reader = new BufferedReader(new FileReader(file));
    String strLine = "";
    boolean selectStartFound = false;
    Pattern pattern = Pattern.compile("<option .*?>(.+?)</option>");
    Matcher matcher = null;
    while ((strLine = reader.readLine()) != null) {
        if (!selectStartFound && strLine.toLowerCase().contains("<select")) {
            selectStartFound = true;
        }
        if (selectStartFound) {
            matcher = pattern.matcher(strLine);
            if (matcher.find()) {
                System.out.println(matcher.group(1));
            }
        }
        if (selectStartFound && strLine.toLowerCase().contains("</select>")) {
            selectStartFound = false;
        }
    }
    reader.close();
}

score 0 · Answer 3 · answered Aug 11 '18 at 08:57

Here's a regex that matches:

/(?:<select.*?>.*?((<option.*?<\/option>\s*)+)<\/select>.*)/ims

By using the s option you can get rid of the mess of trying to match newlines as a special case.

The first capture group captures all of the option tags, the second capture group captures the last option tag. You might need to do this in two steps, processing what the first capture group captures to break it down into all of the separate option tags.

regex101: https://regex101.com/r/lLbsVy/5

Java regex for html select tag parsing

3 Answers3