Returning content of XML element

Question

I'm trying to use xmlstarlet to extract the text in certain elements in this XML feed:

https://services.boatwizard.com/bridge/events/bc0af0c8-4b47-42b3-9a71-5326775344e0/boats?status=on

One of the elements I'd like to extract is the text for city name which is embedded in the XML document as (excluding certain parent elements for clarity):

<Location>
<LocationAddress>
<CityName>St Malo</CityName>
<CountryID>FR</CountryID>
<Postcode>35400</Postcode>
</LocationAddress>
</Location>

I'm trying to extract "St Malo".

I've saved the feed to boats.xml and I've used xmlstarlet el -v boats.xml to figure out the correct XPath name, which seems to be:

ProcessVehicleRemarketingDataArea/VehicleRemarketing/VehicleRemarketingBoatLineItem/Location/LocationAddress/CityName

I am trying the following syntax to extract the text:

xml sel -t -m "ProcessVehicleRemarketingDataArea/VehicleRemarketing/VehicleRemarketingBoatLineItem/Location/LocationAddress/CityName" -v "." -n boats.xml

Have tried many different syntax variations with no success. Almost think it might be the XML file that's off? How I extract "St Malo"?

Answer 1

The XML in the link you supplied declares a default namespace in the VehicleRemarking tag:

 <VehicleRemarketing xmlns="http://www.starstandard.org/STAR/5" ...>

That means you have to declare it with a prefix which you should use to qualify each step of your XPath expression that is part of that namespace:

xml sel -N ns=http://www.starstandard.org/STAR/5 
        -t -m "ProcessVehicleRemarketingDataArea/ns:VehicleRemarketing//ns:CityName" 
        -v "." -n boats.xml

The first element is not part of the namespace, but ns:VehicleRemarketing and all its children are. You can also use just //ns:CityName for the expression, in this case (considering the example you posted - it will return all CityName elements in the file).

Answer 2

You do not actually need a template match (the -m option), because the query is so simple in your case. Taking as input the portion of the XML document that you actually show in the question, using

$ xml sel -t -v "//CityName" -n boats.xml

will result in

St Malo

If the input document actually has namespaces, consider using

$ xml sel -t -v "//*[local-name() = 'CityName']" -n boats.xml

Or, even better, declare this namespace URI together with a prefix, see the answer by helderarocha.