Using Nokogiri with multiple search elements - ruby

Question

In this XML snippet I need to replace the data in the UID for some of the blocks. The actual file contains more than 100 similar blocks.
Although I have been able to extract subsets based on name="Track (Timeline)", I am struggling to reduce this subset to the specific block I need by also using the data in the <TrackID>, if name="Track (TimeLine)" and the text of <TrackID> is 0x1200 then set UID to xxxx.
I am new to Nokogiri and, although I write test scripts, I do not consider myself a programmer.
<StructuralMetadata key="06.0E.2B.34.02.53.01.01.0D.01.01.01.01.01.3B.00" length="116" name="Track (TimeLine)">
<EditRate>25/1</EditRate>
<Origin>0</Origin>
<Sequence>32-04-25-67-E7-A7-86-4A-9B-28-53-6F-66-74-65-6C</Sequence>
<TrackID>0x1200</TrackID>
<TrackName>Softel VBI Data</TrackName>
<TrackNumber>0x17010101</TrackNumber>
<UID>34-C1-B9-B9-5F-07-A4-4E-8F-F4-53-6F-66-74-65-6C</UID>
</StructuralMetadata>
<StructuralMetadata key="06.0E.2B.34.02.53.01.01.0D.01.01.01.01.01.3B.00" length="116" name="Track (TimeLine)">
<EditRate>25/1</EditRate>
<Origin>0</Origin>
<Sequence>35-12-2D-86-E6-74-0B-4C-B4-24-53-6F-66-74-65-6C</Sequence>
<TrackID>0x1300</TrackID>
<TrackName>Softel VBI Data</TrackName>
<TrackNumber>0x0</TrackNumber>
<UID>37-0C-80-34-4C-8D-CE-41-85-F3-53-6F-66-74-65-6C</UID>
</StructuralMetadata>

Answer 1

Using xpath:
//StructuralMetadata
will select all StructuralMetadata elements in your XML. The double slash at the start means to select nodes wherever they appear in the document.
You don't want all the nodes though, you can filter the ones you want with a predicate:
//StructuralMetadata[#name="Track (TimeLine)" and TrackID="0x1200"]
This will select all StructuralMetadata elements that have a name attribute with the value Track (TimeLine), and a TrackID child element with contents 0x1200.
As you're interested in the UID element, you can further refine the expression:
//StructuralMetadata[#name="Track (TimeLine)" and TrackID="0x1200"]/UID
This expression will match all the UID elements that are children of StructuralMetadata elements that match the predicate described above.
Putting this to use:
require 'nokogiri'
# Parse the document, assuming xml_file is a File object containing the XML
doc = Nokogiri::XML(xml_file)
# I'm assuming there is only one element in the document that matches
# the criteria, so I'm using at_xpath
node = doc.at_xpath('//StructuralMetadata[#name="Track (TimeLine)" and TrackID="0x1200"]/UID')
# At this point, doc contains a representation of the xml, and node points to
# the UID node within that representation. We can update the contents of
# this node
node.content = 'XXX'
# Now write out the updated XML. This just writes it to standard output,
# you could write it to a file or elsewhere if needed
puts doc.to_xml

Answer 2

A great way to approach this problem is with the ‘map reduce’ style of programming, which works to take a large list of things and narrow it down and combine it into the result you're after. Specifically, Array#find and Array#select are really useful for this sort of problem. Check out this example:
require 'nokogiri'
xml = Nokogiri::XML.parse(File.read "sample.xml")
element = xml.css('StructuralMetadata').find { |item|
item['name'] == "Track (TimeLine)" and item.css('TrackID').text == "0x1200"
}
puts element.to_xml
This little program first uses the CSS selector to get all of the <StructuralMetadata> elements in the document. It returns an array, which we can filter to just what we want using the Array#find method. Array#select is its cousin which returns an array of all the matching objects instead of the first one it happens to find.
Inside the block we have a test to check if the <StructuralMetadata> tag is the one we’re after. Then it puts the element.to_xml string to the console so you can see which thing it found if you run this as a command-line script. Now you can find the element, you can modify it in the usual way and save out a new XML file or whatever.