Question

I have a column that contains html details in it and within each column, the html contains different http links. I need to find out all the http links in each column.

e.g.: column 1 Row 1

html 
... 
a href = http://www.column1.com....... 
img src=http://www.pic1.com/images/im.jpg...
...
/html

column 1 Row 2

html 
...
a href = http://www.column2.com.......  
img src="http://www.pic2.com/images/im.jpg".... 
/html

in the result I need to get the following list :

• link1 in first column1 in href
  domain in the first column img link
• link2 in second column2 in href
  domain in the second column img2 link

Could anyone help me to find this as i don’t have a clue at all what to do and I’m not good using sql.

Answer 1

well you can use charindex to try and find index of http:// and then you need to find end of URL (it depends on you data, space or “).

you can also write CLR scalar function, implementing regular expression find

using System;
using System.Collections;
using System.Text.RegularExpressions;
using System.Data.SqlTypes;
using Microsoft.SqlServer.Server;

public class CLR
{
    [SqlFunction(FillRowMethodName = "FillRow")]
    public static IEnumerable RegexMatch(string pattern, string text)
    {
        var r = new Regex(pattern);
        return r.Matches(text);
    }

    public static void FillRow(Object obj, out SqlInt32 index, out SqlString match)
    {
        var m = (Match)obj;
        index = new SqlInt32(m.Groups[0].Index + 1);
        match = new SqlString(m.Groups[0].Value);
    }
}

Then you need to create assembly on your SQL Server from this class library

create assembly CLR from 'C:\CLR.dll' with permission_set = safe

Then you can create function

create function RegexMatch(@pattern nvarchar(4000), @text nvarchar(4000))
returns table ([index] int, match nvarchar(4000))
as external name CLR.CLR.RegexMatch