The jsonpath sift will extract JSON elements matching the provided JSON path.

In case the returned data is expected to return JSON, then the JSONPath sift can be used to extract a subtree or a single value. For example, suppose we would like to retrieve the value of the extract key in the Wikipedia query https://en.wikipedia.org/w/api.php?format=json&action=query&prop=extracts&exintro&explaintext&redirects=1&titles=Homie.

Were we to issue the HTTP command in order to retrieve a summary of the article named Homie:

        llInstantMessage(CORRADE, 
            wasKeyValueEncode(
                [
                    "command", "HTTP",
                    "group", wasURLEscape(GROUP),
                    "password", wasURLEscape(PASSWORD),
                    "action", "get",
                    "type", "text",
                    "URL", wasURLEscape("https://en.wikipedia.org/w/api.php?format=json&action=query&prop=extracts&exintro&explaintext&redirects=1&titles=Homie"),
                    "sift", wasListToCSV([
                            "JSONPath",
                            wasURLEscape("$.query.pages.*.extract")
                    ]),
                    "callback", wasURLEscape(URL)
                ]
            )
        );

then the sift JSONPath with the value $.query.pages.*.extract would return the page extract in the data key.

• As one might have guessed, a JSON path is the equivalent of XPath for XML but for JSON. Unfortunately, JSONPath has not been formalized and officially a JSON path is not part of the JSON standard. Nevertheless, since JSON is a well-structured formatting language that, just like XML, can be expressed as a tree structure, the concept of a JSON path remains entirely valid. As will all things that have not yet been formally standardized, be weary of using this feature, and, if there is an easy alternative, then it should be preferred.