Ruby: Extract and operate on partially extracted Nokogiri objects - ruby

Question

require 'nokogiri'
xml = DATA.read
xml_nokogiri = Nokogiri::XML.parse xml
widgets = xml_nokogiri.xpath("//Widget")
dates = widgets.map { |widget| widget.xpath("//DateAdded").text }
puts dates
__END__
<Widgets>
<Widget>
<Price>42</Price>
<DateAdded>04/22/1989</DateAdded>
</Widget>
<Widget>
<Price>29</Price>
<DateAdded>02/05/2015</DateAdded>
</Widget>
</Widgets>
Notes:
This is a contrived example I cooked up as its very inconvenient to post the actual code because of too many dependencies. Did this as this code is readily testable on copy/paste.
widgets is a Nokogiri::XML::NodeSet object which has two Nokogiri::XML::Elements. Each of which is the xml fragment corresponding to the Widget tag.
I am intending to operate on each of those fragments with xpath again, but use of xpath query that starts with // seems to query from the ROOT of the xml AGAIN not the individual fragment.
Any idea why its so? Was expecting dates to hold the tag of each fragment alone.
EDIT: Assume that the tags have a complicated structure that
relative addressing is not practical (like using
xpath("DateAdded"))

Answer 1

.//DateAdded will give you relative XPath (any nested DateAdded node), as well as simple DateAdded without preceding slashes (immediate child):
- dates = widgets.map { |widget| widget.xpath("//DateAdded").text }
# for immediate children use 'DateAdded'
+ dates = widgets.map { |widget| widget.xpath("DateAdded").text }
# for nested elements use './/DateAdded'
+ dates = widgets.map { |widget| widget.xpath(".//DateAdded").text }
#⇒ [
# [0] "04/22/1989",
# [1] "02/05/2015"
#]