Is it possible to serialize a whitelisted subset of a POJO's properties (where the whitelist is known only at runtime) using Jackson?
All the solutions I know of so far (Views, #JsonIgnoreProperties etc.) are static, compile-time solutions.
Further, my backend returns results in the following format:
{
"outcome": "SUCCESS", // an enum
"message": "Success.", // a message for the developer
"result": {
// Some result that's different for each call
}
}
So I am looking for a solution that can be applied to only parts of the object graph (like the contents of the result property).
You probably want to look at #JsonFilter.
See this tutorial on serializing only fields that meet some criteria which includes details of this, and a couple of other methods.
For completeness
#JsonFilter("pojo-filter")
class Pojo {
public int foo;
}
FilterProvider filters = new SimpleFilterProvider()
.addFilter("pojo-filter", new SimpleBeanPropertyFilter() {
#Override
protected boolean include(PropertyWriter writer) {
return "foo".equals(writer.getName())
? Random.nextBoolean()
: true;
}
});
new ObjectMapper().writer().filters(filters).write(new Pojo());
Globally you can use ObjectMapper.setFilterProvider
Related
I'm using Jackson and MongoDB in Java. I want to serialize part of an object as "normal JSON" with Jackson, while a specific property should be serialized as "extended JSON". The reason being to keep the data types of that particular part of the structure intact (i.e. not lose info on e.g. Double vs Int64 vs Int32).
For example, if I have a class:
public class ExampleClassToBeSerialized {
#JsonProperty("itemList")
public List<MyObjectIntendedAsExtendedJson> itemList;
#JsonProperty("something")
public boolean something;
#JsonProperty("somethingElse")
public long somethingElse;
}
public class MyObjectIntendedAsExtendedJson {
public long thisShouldBeExtendedJson;
}
Here I'd like the outer object with something and somethingElse to be plain JSON, while the serialization of itemList should be done as extended JSON, again, to keep datatypes since it might again be used in MongoDB. I'd expect a result similar to this:
{
"itemList": [{
"thisShouldBeExtendedJson": { "$numberLong": "66666666666" }
},{
"thisShouldBeExtendedJson": { "$numberLong": "77777777777" }
}],
"something": false,
"somethingElse": 123123123123
}
I'm unsure how to achieve this particular result. I've managed to get e.g. the itemList as strings which contain extended JSON by using a StdSerializer<MyObjectIntendedAsExtendedJson> and creating e.g. a Document d = new org.bson.Document() and calling d.toJson(JsonWriterSettings.builder().outputMode(JsonMode.EXTENDED).build()). As mentioned, this ends up being an extended JSON object inside a string. I'd like it to be part of the overall string representing the total JSON. In my example only the list would be used for MongoDB interaction, and therefor only that part needs to be extended JSON.
Is there any way to achieve my particular case of part plain JSON, part extended JSON?
I have a web project with 2 Java Entities(Lets Say E1,E2) like how mybatis and VO works.
Object structure:
class E1{
String a;
.... n other data members
E2 e2;
}
class E2{
String b;
.... n other data members
}
Is it possible to make a single class in Android project, i.e.
class E1 {
String a;
String b; //This Data member belongs to class E2
}
and parse it with the help of a framework (like Jackson) or I have to write a custom class for that?
My JSON Data will look like this:
{
"E1": {
"a": "some data",
.
.
.
"E2": {
"b": "some data",
.
.
other data
}
}
}
Is there any API which can do this?
I asked this because with my web Application its not just 2 Class but atleast 10 interconnected class and I am not Using them in my android app. So don't wanna replicate the same classes in android app.
Also if you can suggest any other possible way.
It would be a very bad design practice/approach, making things very difficult to debug, error prone and not future proof (think about it, what if you add to one of the 10 classes a field that conflict with another class' field?).
Anyway, if you still want to trick your way out of the correct approach that would be to have 10 classes, I am not aware of any lib that provides you with this feature. You could parse json ==> 10 Java Map, then merge the 10 Map through the Map::putAll method and finally pass the obtained Map that contains all the objects to Jackson.
Your best bet is to use #JsonAnySetter annotation from Jackson library on the receiver POJO.
public class E1{
public String a;
private Map<String, Object> paramMap = new HashMap<>();
#JsonAnyGetter
public Map<String, Object> getParamMap() {
return paramMap;
}
#JsonAnySetter
public void setParamMap(String s, Object o) {
paramMap.put(s, o);
}
}
This will put every unimplemented attributes in a HashMap.
In combination with #JsonAnyGetter, the serialization of the receiver POJO will give the same result as the JSON input.
Currently my applications can do as follows. It can read a JSON configuration such as:
{
"a": 5,
"b": 3
}
Into a POJO that looks like:
public class AddConf {
private Number a;
private Number b;
// constructor, getters and setters
public int add() {
return a.intValue() + b.intValue();
}
}
Then we can call the add function a return a result of 8 in this example. I would like to augment com.fasterxml.jackson.databind somehow to allow my JSON to have placeholders. So let's say when my configuration I don't know what value b will be until runtime, then I might make a configuration like this:
{
"a": 5,
"b": $b_placeholder$
}
To signify that b value will be provided at runtime. Obviously the above is not a valid JSON, and by default jackson throws an Exception (as it should) when it attempts to parse this with ObjectMapper's readValue. Ideally, I would like to read the above "JSON" (or something equivalent) into a POJO that looks something like this:
public class AddConf {
private Map<String,String> usedPlaceholders;
private Number a;
private Number b;
// constructor, getters and setters
public int add(Map<String,String> runtimeConf) {
if (usedPlaceholders.contains("a")) { // if "a" was a placeholder
a = runtimeConf.get(usedPlaceholder.get("a"));
}
if (usedPlaceholders.contains("b")) { // if "b" was a placeholder
b = runtimeConf.get(usedPlaceholder.get("b"));
}
return a.intValue() + b.intValue();
}
public void setPlaceholder(String key, String value) {
usedPlaceholder.put(key, value);
}
}
The idea is when deserializing the augmented JSON above, it would call setPlaceholder("b", "b_placeholder") instead of setting the value b and therefore when add is called, it will use values in the passed in runtimeConf Map instead of values from the JSON to do its configuration.
Given this, I have 2 questions:
Is there an easier way to accomplish my goal of having "placeholders" in my JSON configuration? It seems if I was to implement my idea here, I would have to override some of the Jackson classes. I would have to override the com.fasterxml.jackson.core.JsonParser to allow $ as a valid token in some scenarios, I would also have to write custom derserializers for all my configuration POJO (such as AddConf). This would likely casade into me having to override much of the jackson code base, which I would rather not do.
If I were to take this approach to override some of the default jackson classes, how might I go about doing that?
I am currently using jackson 2.6.0
You may be looking for annotation #JsonRawValue, use of which allows specifying EXACT contents to include as value, while still taking care of adding necessary separators.
I'm currently trying to use Flexjson to deserialize a JSON String and map it to the Object model of my
Android App. The application is a kind of library with several vendors that can have some catalogs with more catalogs
and documents in them. The json is fetched from a web service I have no influence on and looks something like this:
{
"library":{
"vendor":[
{
"id":146,
"title":"Vendor1",
"catalog":[
{
"id":847,
"document":[
{
"id":1628,
"title":"Document",
...
},
{
...
}
],
"title":"Catalog ",
},
{
...
}
]
},
{
...
}
]
}
}
So each vendor, catalog, document is represented by a JSONObject and all child catalogues and documents are within a JSONArray.
So far everything works fine with Flexjson and the following deserialization code:
LibraryResponse response = new JSONDeserializer<LibraryResponse>()
.use(Timestamp.class, new TimestampObjectFactory())
.deserialize(getLocalLibrary(), LibraryResponse.class);
return response.library;
I do have a Library object that has a List<Vendor>. Each vendor has a List<Catalog> and a List<Document>.
But unfortunately, the web service straps the JSONArrays to simple JSONObjects if a catalog contains only a single document
or a catalog contains just one catalog. So the json in that case looks like this:
"document":
{
"id":1628,
"title":"Document",
...
}
Now Flexjson doesn't know how to deserialize and I end up with a library.vendorX.getDocument() being a List<HashMap> instead of a List<Document>.
One idea is to tell Flexjson explicitly how to handle such cases, but I have no idea where to start with this. Another way could be to parse the initial json manually and replace such JSONObjects with the appropriate JSONArray. But I think that way is not really nice to go, as the library can be pretty deep.
I hope you can provide some guidance here.
Yikes this is some gnarly json mapping going on. What backend coder did that?! #NotHelping.
Well from looking at the code, Flexjson is coded to handle this out of the box. But it looks like it's not passing the typing information down to the bind so it doesn't know what type it's binding into so it just returns a Map. That's a bug that should probably be fixed. Good news is there is a work around.
Anyway, the simplest thing I can think of is to install an ObjectFactory on that list. Then you can check and see if you get a Map or a List when the stream is deserialized. Then you can wrap it in a List and send it on to the appropriate decoder. Something like:
LibraryResponse response = new JSONDeserializer<LibraryResponse>()
.use(Timestamp.class, new TimestampObjectFactory())
.use("library.vendor.values.catalog.values.document", new ListDocumentFactory() )
.deserialize(getLocalLibrary(), LibraryResponse.class);
Then
public class ListDocumentFactory implements ObjectFactory {
public Object instantiate(ObjectBinder context, Object value, Type targetType, Class targetClass) {
if( value instanceof Collection ) {
return context.bindIntoCollection((Collection)value, new ArrayList(), targetType);
} else {
List collection = new ArrayList();
if( targetType instanceof ParameterizedType ) {
ParameterizedType ptype = (ParameterizedType) targetType;
collection.add( context.bind(value, ptype.getActualTypeArguments()[0]) );
} else {
collection.add( context.bind( value ) );
return collection;
}
}
}
}
I think that's roughly what would fix that bug, but should also fix your problem.
I'm using Jackson 1.9.5 in an Android project to parse JSON files.
So far I haven't had any problems, and can parse files fine using the following code:
AssetManager mgr = getAssets();
ObjectMapper mapper = new ObjectMapper();
try {
InputStream ifp = mgr.open("detail_schema.json");
schema = mapper.readValue(ifp, DetailSchema.class);
} catch (IOException e) {
e.printStackTrace();
}
Where the DetailSchema class consists of a mix of primitive types and classes. I'm now running into a problem where I want to parse some JSON like the following:
"fields": {
"Suburb": "Paddington",
"State": "NSW",
"Post Code": "2074",
"Lollipop": "Foo Bar Haz"
}
Where I can't possibly know the map keys before hand (they can be user-defined). As such, I'm not sure what the associated Java class should look like.
Ie, for this example, it could look like:
public class MyClass {
public String Suburb;
public String State;
public String PostCode;
public String Lollipop;
}
But this may not be correct for another instance of the JSON file. Ideally I need some way for Jackson to map values to something like a NameValuePair. I suspect that the automatic object mapping may not be an option in this case - can someone confirm or deny this?
You have two options. Either you can use readTree in ObjectMapper, which returns a JsonNode. Working with a JsonNode is much like working with a tree, so you can get children nodes, read values, et cetera:
InputStream ifp = mgr.open("detail_schema.json");
JsonNode root = mapper.readTree(ifp);
JsonNode fields = root.get("fields");
for (JsonNode children : fields) {
// ...
}
Then you'd need to build your DetailSchema object manually.
Or, you can let Jackson deserialize it as a Map, in which case you'd use your code but where MyClass would be like this:
public class MyClass {
public Map<String, Object> fields;
// getter/setters
}
You can probably type the map values as String as well if you are sure the inputs are text in json. (Actually, I'm not sure what type enforcement Jackson does, maybe it will allow anything anyway...)