Parsing xhtml using java.util.regex

Question

I am parsing an XHTML file using the java.util.regex package and I am perplexed at why the following doesn�t work.
The lines I wish to match are either like this:
<span class="someclass"><b>Some String.</b></span></td>
or
Some String.</td>
The code I use to try to achieve this is:
Pattern somePattern = Pattern.compile(".*(<span class=\"someclass\"><b>)?(.*)[.](</b></span>)?</td>.*");
String s = null;
while((s = br.readLine()) != null) {
if(somePattern.matcher(s).matches()) {
System.out.println("0:" + eventMatcher.group(0));
System.out.println("1:" + eventMatcher.group(1));
System.out.println("2:" + eventMatcher.group(2));
System.out.println("3:" + eventMatcher.group(3));
I expect to get as output
0:<span class="someclass"><b>Some String.</b></span></td> 1:<span class="someclass"><b>
2:Some String
3:</b></span>
or
0:Some String.</td>
1:null
2:Some String
3:null
depending on which lines provide the match as mentioned above. Instead I get:
0:<span class="someclass"><b>Some String.</b></span></td>
1:null
2:(empty string)
3:</b></span>
or
0:Some String.</td>
1:null
2:(empty string)
3:null
Any ideas? Thanks in advance.

Answer 1

Consider the terms of ".*(<span class=\"someclass\"><b>)?(.*)[.](</b></span>)?</td>.*"
.* - greedily collect characters
(<span class=\"someclass\"><b>)? - optionallly collect information taht will always be matched by the previous .* pattern so will be empty.
(.*) - greedily collect characters that will also have been swallowed by the first .* so will be empty
[.] - a single .
(</b></span>)? - optionally collection
</td> - must be there
.* - collect the rest of the charcters in the line.
Therefore in general groups 1 and 2 will be empty because the first .* will have collected the information you wanted to capture!
You could just make the first .* non-greedy by using .*? but this may fail for other reasons.
So, in general terms, what are you trying to extract?