Question

[the JSoup discussion page suggested I ask my question here.]

So, I am no regex expert, but I’m wondering about the results I get from the jsoup
getElementsByAttributeValueMatching() method.

If I have an html page that has (among others) the following links:

<a href="/tweb/tiles/twr/EIDS_AT_20130108T134335/01/">Parent Directory</a>
<a href="1357681618315/">1357681618315/</a>
<a href="1357681649996/">1357681649996/</a>

And I query with:

Elements dirs = baseDir.getElementsByAttributeValueMatching("href", Pattern.compile("[0-9]+/"));

hoping to get just the 2 links that have only numbers (and a slash at the end.)

However, I get all 3 links back.

I wrote a quick test program to check java’s Pattern Matcher response to that regex with the 3 href strings, and it returns only the two with only numbers as I would expect:

String a = "/tweb/tiles/twr/EIDS_AT_20130108T134335/01/";
String b = "1357681618315/";
String c = "1357681649996/";

Pattern p = Pattern.compile("[0-9]+/");

System.out.println("a:"+ p.matcher(a).matches());
System.out.println("b:"+ p.matcher(b).matches());
System.out.println("c:"+ p.matcher(c).matches());

returns:
a:false
b:true
c:true

So, my question is, what am I missing?

thanks,
Linus

Answer 1

Jsoup uses Matcher#find(), not Matcher#matches(). So, you need to supply ^ and $ yourself.

Elements dirs = baseDir.getElementsByAttributeValueMatching(
    "href", Pattern.compile("^[0-9]+/$"));

Here are javadoc extracts of relevance explaining the difference (emphasis mine):

find

…

Returns:

true if, and only if, a subsequence of the input sequence matches this matcher’s pattern

matches

…

Returns:

true if, and only if, the entire region sequence matches this matcher’s pattern

As to why Jsoup uses find() instead of matches(), that’s a question you’ve to ask to its creator.