Which approach for injecting urls in text?

I have a task using c#. It's part of an existing web forms project.

What I want to do is:

• inside the page's <main></main> tag (which has a runat="server"),
• find the <a></a> tags which have an attribute containing the page-id
  (E.g. data-page-id="12345");
• fetch the fully qualified urls from the database for each PageId; // the easy bit
• inject the urls back into the text/controls (href="mysite.com/an-seo-worthy-url").

Where I'm struggling is with the initial approach. Do I use the FindControls() functionality?; Regex? HtmlAgilityPack? Is any one of these preferred for this kind of task, or is there another way?

Anyone else done something similar, that can point me in a particular direction?

Ok, so the goal is to have hyperlinks on a page, but the resolution of the link is to occur say from a database.

I would NOT replace all of the hyperlinks on page load, since this is too high of a database cost. Say we had 10 links on a page. If we replace the links at page load time, then that is 10 database operations. That is rather expensive, since say if we had 20 users on the site, then we done 200 database operations. That is simply WAY TOO many database hits and pulls. This is especially the case since the user will NOT YET have clicked on any of the links!

So, I suggest using and dropping into the page a SERVER SIDE hyper-link control.

Say, like this (2 example link buttons)

        <asp:LinkButton ID="L1" runat="server"
            OnClick="mylinkresolve_Click"
            data-page-id="12345"                
            >My Link Text</asp:LinkButton>

        <br />

        <asp:LinkButton ID="L2" runat="server"
            OnClick="mylinkresolve_Click"
            data-page-id="6789"                
            >My Link Text2 </asp:LinkButton>

Now, you could consider using the ID in above, but either way, don't really matter. For above, we use data-page-id.

So, now in the code behind, we have ONE routine that ALL link buttons point to.

The code would be like this:

    protected void mylinkresolve_Click(object sender, EventArgs e)
    {
        LinkButton btn = (LinkButton)sender;
        string strSQL = @"SELECT Hlink FROM tblMyLinks 
                        WHERE LinkID = @LinkID";

        SqlCommand cmdSQL = new SqlCommand(strSQL);
        cmdSQL.Parameters.Add("@LinkID", SqlDbType.NVarChar).Value = btn.Attributes["data-page-id"];

        DataTable rstLink = General.MyRstP(cmdSQL);
        if (rstLink.Rows.Count > 0)
        {
            string jumpURL = rstLink.Rows[0]["URLLink"].ToString();
            Response.Redirect(jumpURL);
        }
    }

So, based on the link button clicked, we get the data-page-id, pull the value from the database and then jump to that URL.

With the above design, then it don't matter if we have 1, or 20 links on the page, since at page load time, we don't do ANY database operations, but ONLY when a user actually clicks on a button.

As a result, we don't incur any database pull and operations at page load time, and thus will not place unnecessary load on the server.

Note how each LinkButton can use/call the ONE SAME routine in code behind.