Accessing Information from XML/JSON using XPath/Pointer

[Morten MacFly]

Since version 22.01, WCM supports to extract information from XML and JSON documents. These can be numbers (for example) that be further processed using the interpreter function for a comparison against an expected / changed value to create an web change alert. Furthermore, if the server features a REST API it can be used to access dedicated information using dedicated server-requests.

• For XML, WCM supports the XML/XPath notation to do so as implemented in the pugixml library
• For JSON, JSON/Pointer and JSONPath are supported as implemented in the jsoncons library

How does it work? Consider this document for XML:

<books>
  <book>
        <title>A Wild Sheep Chase</title>
        <price>22.72</price>
  </book>
  <book>
        <title>The Night Watch</title>
        <price>23.58</price>
  </book>
  <book>
        <title>The Comedians</title>
        <price>21.99</price>
  </book>
</books>

...then you can access the price of the second book (the index is zero-based!) by this XML/XPath notation:
/books/book[1]/price
...which will return: 23.58

For JSON it is similar. Consider this JSON document:

{
    "books":
    [
        {
            "title" : "A Wild Sheep Chase",
            "price" : 22.72
        },
        {
            "title" : "The Night Watch",
            "price" : 23.58
        },
        {
            "title" : "The Comedians",
            "price" : 21.99
        },
    ]
}

...then you can access the price of the second book by this JSON/Pointer notation:
/books/1/price
...alternatively you can use JSONPath like this:
$.books.1.price

Again, if you let an interpreter operation follow then you track prices of items, for example.

For XML, there is something special to consider:
XML/XPath does not work on HTML because HTML is not a valid XML document you can apply XPath on. For these cases, WCM offers the option to try to convert HTML to a valid XML document (in fact, it becomes something like XHTML). You have to enable this conversion in the options of the XML/XPath filter. On the other hand, if the webpage is already XML or XHTML you don't need and should not apply the conversion as it may lead to errors otherwise.

[Gitoffthelawn]

Is there a way to select the price of "The Night Watch" if its position in the list may change?

For example, something like: /books/title:"The Night Watch"/price

If not, this functionality will be exceptionally useful to add!

[Morten MacFly]

Good question. Honest answer: I don't now. I am using the syntax of the JSONCons library (https://danielaparker.github.io/jsoncons/) which ships with many examples (on that homepage and also here: https://github.com/danielaparker/jsoncons/tree/master/examples/src and here: https://github.com/danielaparker/jsoncons?tab=readme-ov-file#E1). You might want to check if you find something useful there for your purpose and then I can check if there is a need to change / enhance the implementation.

[GABRIEL OLIVEIRA DINIZ]

Thank you for this update! I think it has a little bit relation with my question, right? Anyway, thank you!

[Morten MacFly]

Yes, indeed - it could probably solve your problem! For me it was more to support REST APIs on servers where you send a query and get often JSON in return. Glad to hear it may be of further use...