For a freshly loaded SERP page, your Xpath should return the destination URL for each SERP listing. It appears as though it's extracting Google's tracking URL instead.
@href value with an internal Google URL that tracks the click before redirecting you to the desired page. The JS function also adds a new attribute called
@data-href that contains the original URL. You can visualize this by hovering over a link, then comparing the URL shown in the status bar before and after right-clicking the link. Here's what happens to the
Same listing, After click:
Normally, the tracking URL begins with
/url?.... I believe the
/search?... instance you're getting may be the Related Searches at the bottom of the SERP?
Hopefully this is the right track.