What is meant by deserializing json data.? What is the need to do it? I have been told to do this in my project when given a json object like this:
###format of a json record in a json file with .json extension
####
'''{
"firstname":"stack",
"lastname":"overflow",
"age":xx,
"height":x.y,
"phonenumbers":[
{
"type":"home","number":"xx-xx.xxx.x.x.x"
},
{
"type":"fax","number":"x.x.x.x.x,x,x.x.x.x"
}
]
}
'''
defining a class object
It means converting a JSON string into an object in your programming language, such as a Java POJO, a Ruby hash, a JavaScript object, etc.
The purpose is to allow your programming language to read and manipulate the data contained therein.
Related
I am looking at extracting the root element of a JSON document. It looks like this is possible neither using JsonPointer nor JsonPath as my attempts to look up for such an expression has been unsuccessful. Any tips would be appreciated. TIA.
Sample document:
{
"MESSAGE1_ROOT_INPUT": {
"CTRL_SEG": "test"
}
}
The below using gson 2.9.0:
$.*~
produces:
{"CTRL_SEG": "test"}
while JSONPath Online produces this:
[
"MESSAGE1_ROOT_INPUT"
]
The attempt is to get text "MESSAGE1_ROOT_INPUT" using JsonPath/JsonPointer expression(s). Note that, extracting this the traditional (substring or regex on a stringified json text) way, would preferably be my last resort.
Background: We are building an API service that accepts JSON documents with different roots. Such as, MESSAGE2_ROOT_INPUT, MESSAGE3_ROOT_INPUT, etc. It is based on this, the routing of a message further will occur.
Supported/Employed Languages: Java/GSON Library/RegEx
Gson does not natively support JSONPath or JSON Pointer. However, you can quite efficiently obtain the name of the first property using JsonReader:
public static String getFirstPropertyName(Reader reader) throws IOException {
// Don't have to call JsonReader.close(); that would just close the provided reader
JsonReader jsonReader = new JsonReader(reader);
jsonReader.beginObject();
return jsonReader.nextName();
}
There are however two things to keep in mind:
This only reads the beginning of the JSON document; it neither verifies that the complete JSON document has valid syntax, nor checks if there might be more top-level properties
This consumes some data from the Reader; to further process the data you have to buffer the data to allow re-reading it again (you can also first store the JSON in a String and pass a StringReader to JsonReader)
For a json array like this:
[
my-json-obj1,
my-json-obj2,
my-json-obj3,
....
my-json-objN
]
And MyJsonObj class that represents a mapping of single object in array I can say:
val myJson = '''[...]'''
ZStream
.fromIterable(myJson.toSeq)
.via(JsonDecoder[MyJsonObj].decodeJsonPipeline(JsonStreamDelimiter.Array))
to parse that array in a "streaming" way, i.e. emit mapped objects as they are parsed form the input as opposed to reading all the input first and then extracting the objects.
How can I do the same if the array is nested inside a json object say like this?:
{
"hugeArray":
[
my-json-obj1,
my-json-obj2,
my-json-obj3,
....
my-json-objN
]
}
I trawled through zio-json source code, but I can't find any foothole there for this use case. I guess I could carve out that array from the json document and feed that to decodeJsonPipeline. Is there any better, json-syntax aware way of doing this? If not directly in zio-json perhaps with help of some other open source json libraries?
I need to build up a JSON node in XQuery in MarkLogic. I know that I can use xdmp:unquote() to parse from a string into a node(). However, I’d like to build the JSON programmatically, without ugly string concatenation. I can use computed element constructors to build XML nodes in XQuery. Is there something similar for JSON nodes?
JSON is implemented in MarkLogic as an extension to the XML data model. MarkLogic 8 introduces object-node, array-node, number-node, boolean-node, and null-node tests and constructors. Thus, in XQuery you can build JSON with computed constructors, just like you would with XML. For example,
object-node {
"key" || fn:string(xdmp:random(100)): array-node { 1, 2, 3 },
"another": object-node { "child": text {'asdf'} },
"lastButNotLeast": boolean-node { fn:true() }
}
will create the JSON,
{
"key47": [1, 2, 3],
"another": {
"child": "asdf"
},
"lastButNotLeast": true
}
Aside: In JavaScript you can build JSON-like structures as JavaScript objects using JavaScript syntax. You can convert a JavaScript object into a JSON node using xdmp.toJSON(). Most builtin functions that require a JSON node, however, will do this conversion automatically, such as xdmp.documentInsert().
I have a method in my main controller that return a string that I want to render as JSON.
So I am importing "import grails.converters.JSON" and calling
myMethod() as JSON
, and it works fine. But when I need to get some details of the json response in my integration test.
So in my integration test I have:
void testfoo() {
def bar = controller.myMethod();
def bar.name; //fails
JSON.parse(bar.toString()).name; // doesn't fail
....
..
}
any idea why I need to convert it to a string and then again to a JSON, since it already a JSON?
The value you get back from your method is a grails.converters.JSON, which is not a directly accessible JSON tree as such, but simply an object that knows how to serialize itself as JSON when required. If you want direct access to the JSON tree structure then you need to tell the grails.converters.JSON object to serialize itself and then pass that JSON to JSON.parse to turn it into a JSONElement (or one of its subclasses, in this case presumably a JSONObject).
I have been trying to evaluate GWT Autobean feature to decode/encode JSON object to domain objects for REST calls.
Following the example : http://code.google.com/p/google-web-toolkit/wiki/AutoBean#Quickstart
I was able to convert a singular JSON object to a domain object:
AutoBean<Person> personBean = AutoBeanCodex.decode(factory, Person.class, JsonResources.INSTANCE.json().getText());
where JsonResources.INSTANCE.json() is returning a JSON string.
However, I haven't been successful to convert a list of Person objects from JSON.
It would be helpful, if anyone has an example of this?
Thanks!
Well the only way I can think of is to create a special autobean, which will have List<Person> property. For example:
public interface Result {
void setPersons(List<Person> persons);
List<Person> getPersons();
}
And example json string:
{
persons:[
{"name":"Thomas Broyer"},
{"name":"Colin Alworth"}
]
}
UPDATE:
Workaround when input JSON is an array ( as suggested by persons[0] in comments).E.g. JSON looks like this:
[{"name":"Thomas Broyer"},{"name":"Colin Alworth"}]
And parsing code looks like this:
AutoBeanCodex.decode(factory, Result.class, "{\"persons\": " + json + "}").getPersons();