C# Regular Expression and Replace on part of the match

Question

So I have a regular expression that will look into a string and match all of the relative anchor links, like: Leaderboard

It will not match where the href starts with HTTP or HTTPS.

Expression is:

<a.*?href="([^http]|[^https]).*?"[^<]

That part is good for now.

However, I cannot seem to figure out that after I get the match I want to replace just the href name with routerLink.

This:

<a href="/leaderboard">Leaderboard</a>

Becomes:

<a routerLink="/leaderboard">Leaderboard</a>

Note href is now routerLink.

There are 20+ matches so I can't simply do a replace with Leaderboard I need to keep the relative path the same. Literally the only thing in the matched string that gets replaced is href for routerLink and the value of that attribute stays as is.

This part is giving me issue!

Any ideas here?

Thanks

Your regex is very peculiar and doesn't really do what you think. — NetMage, Aug 25 '17 at 22:20
Just updated my expression. It had been formatted. If that is wrong I can create a post just for that one and update you with the link. — Jordan McDonald, Aug 25 '17 at 22:26
@JordanMcDonald Any specific reason to parse an html with regex? — L.B, Aug 25 '17 at 22:32
@L.B I just have a string of HTML I get from a service and I need to modify it before it gets rendered on the client side. Any better ideas? — Jordan McDonald, Aug 25 '17 at 22:33
@JordanMcDonald you can use *HtmlAgilityPack*. Seems like you haven't read https://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags — L.B, Aug 25 '17 at 22:37
@L.B Thanks I am looking into this now as possible better solution. Thanks for the good looking out! — Jordan McDonald, Aug 25 '17 at 22:44

L.B · Answer 1 · 2017-08-25T23:09:48.537

3

Eventhough other post is already accepted, I want to post an alternative without regex

var doc = new HtmlAgilityPack.HtmlDocument();
doc.LoadHtml(html);

foreach(var a in doc.DocumentNode.SelectNodes("//a[@href]"))
{
    var href = a.Attributes["href"];
    href.Remove();
    a.Attributes.Add("routerLink", href.Value);
}

var newHtml = doc.DocumentNode.OuterHtml;

A Linq query can also be used for the same XPath

foreach (var a in doc.DocumentNode.Descendants("a")
                     .Where(a => a.Attributes["href"] != null))

edited Aug 25 '17 at 23:09

answered Aug 25 '17 at 22:47

L.B

114,136
19
178
224

Yes that is very clean. Thank you – Jordan McDonald Aug 25 '17 at 23:45
Your method is nice but you aren't filtering on only the relative href links, are you? – NetMage Aug 29 '17 at 19:24

NetMage · Accepted Answer · 2017-08-25T22:43:56.653

0

Your pattern doesn't do what you think - see what happens if you have href="halfway" with your pattern. The pattern you have says <a.*?href="([^http]|[^https]).*?"[^<] which breaks down as:

Find the characters <a literally.
Optionally skip as few characters as possible .*? to match
Match the characters href=" literally
Accept either a character that is not one of h, t, t, p or not one of h, t, t, p, s. ([^http]|[^https]) - note that character lists in square brackets [,] represent one character matched.
Optionally skip as few characters as possible .*? to match
Match the character " literally
Match a character that is not <

While this online site won't handle all of .Net regular expressions, it shows the issues and explains how some of the matching operations work: https://regex101.com/r/raoCcA/1

This should work:

var pattern = @"href=(?=""(?!http|https))";

var ans = Regex.Replace(src, pattern, "routerLink=");

You can try to restrict to href inside a tags if necessary, but it starts to get too complicated for regular expressions:

var pattern = @"(?<=<a([^<>]|<!--|-->)+)href=(?=""(?!http|https))";

edited Aug 25 '17 at 22:43

answered Aug 25 '17 at 22:25

NetMage

26,163
3
34
55

Just saw this. Will test this out. – Jordan McDonald Aug 25 '17 at 22:26
With your new pattern you appear to be trying to limit to replacing `href` in `a` tags - is that correct? – NetMage Aug 25 '17 at 22:31
I have a bunch of tags and I only want the ones that do not contain HTTP or HTTPS in the href value. So really only I want relative paths to match like foo – Jordan McDonald Aug 25 '17 at 22:34
But do you care about href in other contexts e.g. HTML comments (`` or in body text `
see href="test" as an example
`)? – NetMage Aug 25 '17 at 22:36
That is a good call. didn't really think of it. Don't think it will be an issue. What you gave me above is working. Do you want me to create a new post to get a more solid and safe Regex? I only care about tags. No other tags need to be matched. – Jordan McDonald Aug 25 '17 at 22:40
I added an expression that attempts to include only `a` tags. It probably isn't possible to create a regular expression to handle all of HTML5 conditions. – NetMage Aug 25 '17 at 22:42

C# Regular Expression and Replace on part of the match

2 Answers2