用于HTML的正则表达式C＃

Question

I have following regex:

^(<span style=.*?font-weight:bold.*?>.*?</span>)

It matches the following code:

<span style="font-family:Arial; font-size:10pt"> r.</span></p><p style="margin:0pt"><span style="font-family:Arial; font-size:10pt; font-weight:bold">&#xa0;</span>

But I would like to match only this part (last span containing font-weight:bold style)

<span style="font-family:Arial; font-size:10pt; font-weight:bold">&#xa0;</span>

Answer 1

Use HTML Agility Pack to parse html:

HtmlDocument doc = new HtmlDocument();
doc.LoadHtml(htmlContent);

var boldSpans = from s in doc.DocumentNode.SelectNodes("//span")
                let style = s.Attributes["style"].Value
                where style.Contains("font-weight:bold")
                select s;

Or even better xpath, which selects all nodes in one line:

doc.DocumentNode.SelectNodes("//span[contains(@style, 'font-weight:bold')]")

Answer 2

Don't use ^ since the line doesn't start with the span you want to match.

<span style=["'][^'"]*font-weight:bold[^'"]*['"]>[^<]*</span>

Or as escaped string:

"<span style=[\"'][^'\"]*font-weight:bold[^'\"]*['\"]>[^<]*</span>"

This matches strings starting with <span style= followed by single or double quote ' , " . Then [^'"]* allows all characters except ending quotes.

Match string font-weight:bold , followed again by any amount of characters except ending qoutes leading up to the real ending qoutes and ending tag: [^'"]*['"]> .

(Note that you might or might not want to allow more attributes before and after the style attribute. In that case you need to alter the regex)

span may contain any amount of any characters except start tag < , then string has to end with closing </span> tag.

Answer 3

remove the ^, because it means beginning of the line. Therefore it will always get the first span. More so because .* means (any characters at all).

doing this the first match may stil be the output you have now, but the second match should be what you're after.

Furthermore tools like regexbuddy and such are good for testing Regex's.