XPath Query: get attribute href from a tag

The answer shared by @mockinterface is correct. Although I would like to add my 2 cents to it.

If someone is using frameworks like scrapy the you will have to use /html/body//a[contains(@href,'com')][2]/@href along with get() like this:

response.xpath('//a[contains(@href,'com')][2]/@href').get()

For the following HTML document:

<html>
  <body>
    <a href="http://www.example.com">Example</a> 
    <a href="http://www.stackoverflow.com">SO</a> 
  </body>
</html>

The xpath query /html/body//a/@href (or simply //a/@href) will return:

    http://www.example.com
    http://www.stackoverflow.com

To select a specific instance use /html/body//a[N]/@href,

    $ /html/body//a[2]/@href
    http://www.stackoverflow.com

To test for strings contained in the attribute and return the attribute itself place the check on the tag not on the attribute:

    $ /html/body//a[contains(@href,'example')]/@href
    http://www.example.com

Mixing the two:

    $ /html/body//a[contains(@href,'com')][2]/@href
    http://www.stackoverflow.com