JSON & Jackson

II. Detailed Study Notes#

One numbered concept each: claim → plain-English build-up → why it matters → traps. + marks my additions and corrections.

1. Persistence: why in-memory data disappears#

Data stored in Java collections lives only in RAM, so it vanishes the moment the program stops; to keep it, you must write it somewhere durable.

When I put objects into an ArrayList, HashMap, or any collection, those objects sit in the JVM's heap, which is just a region of RAM. RAM is volatile: cut the power (or end the process) and it's wiped. So once main() returns, every collection and everything in it is gone.

Persistence means surviving beyond a single program run. To persist data I have to move it out of RAM into something non-volatile: a file on disk, a database, or send it across the network to another machine that stores it. + The technical distinction is volatile (RAM, registers, cache) vs non-volatile (disk, SSD, DB) storage.

2. Why persisted data must be language-neutral#

If you want a program in any language to read your saved data, you can't save raw Java objects, you save a neutral, text-based format.

A Java object's in-memory layout is specific to the JVM. A Python or Go program has no idea what a Java HashMap's bytes mean. So when data needs to cross language (or machine) boundaries, we agree on a common, portable format, usually plain text with a documented structure. Both sides know the format's rules, so both can read and write it.

This is exactly why data on the internet travels as text formats like JSON rather than as Java/Python objects: the receiver might be written in a totally different language.

3. Serialization & Deserialization#

Serialization converts an in-memory object into a portable format; deserialization rebuilds the object from that format. They're encode/decode for data.

Serialization = take a live object and turn it into a flat sequence (a string or byte stream) you can store or send. Deserialization = take that flat sequence and reconstruct the equivalent object in memory. The pair is sometimes written (de)serialization.

The same idea shows up everywhere, even YouTube: video is encoded into data packets to transmit, then a decoder on your device turns packets back into a playable stream. Encoding/decoding is the general name; serialization/deserialization is that idea applied to program objects. + "Marshalling / unmarshalling" is a near-synonym you'll see in other ecosystems.

→ See code: serialize/deserialize round-trip

4. Data file formats: JSON and the field#

JSON is the dominant human-readable format for internet data exchange; binary formats exist for speed and size but you can't read them by eye.

JSON (JavaScript Object Notation) is the most common format for transferring data over the internet. It's plain text, so a human can open and read it, and microservices overwhelmingly use it to exchange information. Beyond JSON there's a whole family:

5. JSON syntax rules#

JSON is built from objects, arrays, and a small fixed set of value types, with strict punctuation.

A JSON document is one value. Usually that value is an object: an unordered set of "name": value pairs wrapped in { }. The pieces:

• Object { }, key-value pairs; nesting allowed (a value can itself be an object/array).
• Pair, each name is followed by a colon : and pairs are separated by commas ,.
• String, text inside double quotes. Keys are always strings.
• Number, integers or decimals, no quotes (e.g. 3925, 25.7).
• Boolean, true or false (lowercase).
• null, a special value meaning "empty" for any valid type.
• Array [ ], an ordered list of supported values.

{
  "firstName": "John",
  "lastName": "Doe",
  "rollNumber": 3925,
  "bodyMassIndex": 25.7,
  "married": false,
  "phoneNumbers": [2405820570, 750285042],
  "address": { "city": "Atlanta", "zipCode": 50005 },
  "middleName": null
}

6. POJO: the Java side of the bridge#

A POJO is a plain Java class of simple fields that Jackson maps to and from JSON.

+ POJO = Plain Old Java Object, an ordinary class with fields (and usually getters/setters), not tied to any framework or special base class. Java holds information as objects, so to (de)serialize I need a class whose shape mirrors the JSON.

public class Member {
    public String name;
    public Integer age;
    public String secretIdentity;
}

Given JSON {"name":"JJonahJameson","age":60,"secretIdentity":"OmniMan"}, Jackson lines up each JSON key with the same-named Java field.

7. Jackson & ObjectMapper#

Jackson is the Java library for JSON; its ObjectMapper is the one object that does both serialization and deserialization.

Jackson is the de-facto JSON library in Java. The workhorse is ObjectMapper. Two directions, two methods:

+ Correction worth noting: writeValue(File, obj) returns void (it writes to the file). If you want the JSON as a return value, use writeValueAsString(obj). The original "String s = om.writeValue(...)" mixes the two, the file overload doesn't return the string.

How does Jackson know which field maps to which key? By name matching: JSON key "age" → field/property named age. + Under the hood Jackson matches against properties, which it discovers from public fields or from getter/setter method names (getAge() → property age).

→ See code: ObjectMapper round-trip

8. Activity 1, parsing stock trades & edge cases#

A real array of trade objects maps to a List<Trade>, and poking at the POJO reveals exactly how Jackson handles mismatches.

Input is an array, so deserialize into a list, not a single object. The questions this activity provokes (and their answers):

→ See code: Trade list parsing + edge experiments

9. Activity 2, @JsonProperty when keys don't match #

When the JSON key and your Java field name differ, @JsonProperty declares the exact key to bind to, no need to rename your fields.

Sometimes the incoming JSON uses keys you can't or don't want to mirror in Java (e.g. "1.symbol", or snake_case keys vs camelCase fields). Renaming Java fields to match ugly keys is bad, so instead annotate:

public class Trade {
    @JsonProperty("1.symbol")
    public String symbol;

    @JsonProperty("2.quantity")
    public int quantity;

    @JsonProperty("3.purchaseDate")
    public String purchaseDate;
}

Now "1.symbol" binds to symbol both ways: on deserialize the key fills the field; on serialize the field is written back out as "1.symbol". + The annotation is bidirectional, it controls both reading and writing.

→ See code: @JsonProperty mapping

10. Activity 2.1, @JsonIgnoreProperties for unknown fields#

When JSON contains fields your POJO doesn't have, Jackson throws by default; @JsonIgnoreProperties(ignoreUnknown=true) tells it to skip them quietly.

If the JSON gains an extra key (say "4.weather":"Rainy") that has no matching field, default Jackson raises UnrecognizedPropertyException and deserialization fails. You rarely want a new upstream field to crash your parser, so:

@JsonIgnoreProperties(ignoreUnknown = true)
public class Trade {
    @JsonProperty("1.symbol")        public String symbol;
    @JsonProperty("2.quantity")      public int quantity;
    @JsonProperty("3.purchaseDate")  public String purchaseDate;
}

The annotation goes on the class. Now unknown keys are ignored and parsing succeeds.

→ See code: @JsonIgnoreProperties

Connections between these ideas#

• Persistence → Serialization: wanting data to survive forces you to serialize it.
• Cross-language → JSON: needing many languages to read it forces a neutral text format.
• JSON → Jackson: in Java, JSON ⇄ object conversion is done by Jackson's ObjectMapper.
• POJO names → annotations: when POJO field names can't match JSON keys, annotations bridge the gap (@JsonProperty) and handle surprises (@JsonIgnoreProperties).

III. Code Samples (runnable)#

Real, runnable Java + Jackson. Each block has a filename, a run command, and expected output. + Setup uses the Jackson Databind dependency; see the Lab for the one-time setup.

A. ObjectMapper round-trip (serialize + deserialize) #

public class Member {
    // public fields so Jackson can see them with no getters/setters
    public String name;
    public Integer age;
    public String secretIdentity;

    // Jackson needs a no-arg constructor to build the object on read
    public Member() {}
    public Member(String name, Integer age, String id) {
        this.name = name; this.age = age; this.secretIdentity = id;
    }
}

import com.fasterxml.jackson.databind.ObjectMapper;

public class Main {
    public static void main(String[] args) throws Exception {
        ObjectMapper om = new ObjectMapper();

        // --- SERIALIZE: object -> JSON string ---
        Member m = new Member("JJonahJameson", 60, "OmniMan");
        String json = om.writeValueAsString(m);   // returns the JSON text
        System.out.println(json);

        // --- DESERIALIZE: JSON string -> object ---
        String input = "{\"name\":\"Peter\",\"age\":29,\"secretIdentity\":\"Spidey\"}";
        Member back = om.readValue(input, Member.class);
        System.out.println(back.name + " / " + back.age);
    }
}

B. Activity 1, parse a list of trades + edge experiments#

public class Trade {
    public String symbol;
    public int quantity;
    public String purchaseDate;

    @Override public String toString() {
        return symbol + " x" + quantity + " @" + purchaseDate;
    }
}

import com.fasterxml.jackson.databind.ObjectMapper;
import com.fasterxml.jackson.core.type.TypeReference;
import java.util.List;

public class ParseTrades {
    public static void main(String[] a) throws Exception {
        String jsonArray = "[{\"symbol\":\"AAPL\",\"quantity\":100,\"purchaseDate\":\"2019-01-02\"}," + " {\"symbol\":\"MSFT\",\"quantity\":10,\"purchaseDate\":\"2019-01-02\"}]";

        ObjectMapper om = new ObjectMapper();
        // TypeReference preserves the generic type List<Trade> at runtime
        List<Trade> trades = om.readValue(jsonArray, new TypeReference<List<Trade>>() {});

        trades.forEach(System.out::println);
    }
}

C. Activity 2, @JsonProperty for mismatched keys#

import com.fasterxml.jackson.annotation.JsonProperty;

public class Trade {
    @JsonProperty("1.symbol")        public String symbol;
    @JsonProperty("2.quantity")      public int quantity;
    @JsonProperty("3.purchaseDate")  public String purchaseDate;

    @Override public String toString() {
        return symbol + " x" + quantity;
    }
}

String json = "{\"1.symbol\":\"AAPL\",\"2.quantity\":100,\"3.purchaseDate\":\"2019-01-02\"}";
ObjectMapper om = new ObjectMapper();

Trade t = om.readValue(json, Trade.class);   // "1.symbol" -> symbol
System.out.println(t);

String out = om.writeValueAsString(t);          // writes "1.symbol" back
System.out.println(out);

D. Activity 2.1, @JsonIgnoreProperties for unknown fields#

import com.fasterxml.jackson.annotation.JsonProperty;
import com.fasterxml.jackson.annotation.JsonIgnoreProperties;

@JsonIgnoreProperties(ignoreUnknown = true)  // skip keys we don't model
public class Trade {
    @JsonProperty("1.symbol")        public String symbol;
    @JsonProperty("2.quantity")      public int quantity;
    @JsonProperty("3.purchaseDate")  public String purchaseDate;
}

// note the extra "4.weather" key that Trade has no field for
String json = "{\"1.symbol\":\"AAPL\",\"2.quantity\":100," + "\"3.purchaseDate\":\"2019-01-02\",\"4.weather\":\"Rainy\"}";

ObjectMapper om = new ObjectMapper();
Trade t = om.readValue(json, Trade.class);  // weather ignored, no crash
System.out.println(t.symbol + " " + t.quantity);

IV. Hands-On Lab (practice)#

Six graded exercises. Predict the output before you run, then verify. Solutions are collapsed.

plugins { id 'application' }
repositories { mavenCentral() }
dependencies {
    // brings databind + core + annotations transitively
    implementation 'com.fasterxml.jackson.core:jackson-databind:2.17.1'
}
application { mainClass = 'Main' }

1. Serialize a single object WARM-UP#

Goal: turn one Member into a JSON string.

1. Create Member with fields name, age, secretIdentity + no-arg constructor.
2. In Main, build a Member and call writeValueAsString.
3. Print the result.

ObjectMapper om = new ObjectMapper();
Member m = new Member("Ada", 36, "Countess");
System.out.println(om.writeValueAsString(m));

2. Deserialize back into an object WARM-UP#

Goal: rebuild a Member from a JSON string and read a field.

1. Take the JSON from Ex 1 as a string literal (escape the quotes).
2. Call readValue(json, Member.class).
3. Print member.name and member.age.

String json = "{\"name\":\"Ada\",\"age\":36,\"secretIdentity\":\"Countess\"}";
Member m = new ObjectMapper().readValue(json, Member.class);
System.out.println(m.name + " / " + m.age);

3. Parse a JSON array into a List CORE#

Goal: deserialize the two-trade array into List<Trade> and sum the quantities.

1. Define Trade (symbol, quantity, purchaseDate).
2. Read with new TypeReference<List<Trade>>(){}.
3. Loop and add up quantity.

String arr = "[{\"symbol\":\"AAPL\",\"quantity\":100,\"purchaseDate\":\"2019-01-02\"}," + " {\"symbol\":\"MSFT\",\"quantity\":10,\"purchaseDate\":\"2019-01-02\"}]";
List<Trade> ts = new ObjectMapper()
    .readValue(arr, new TypeReference<List<Trade>>(){});
int total = ts.stream().mapToInt(t -> t.quantity).sum();
System.out.println("total = " + total);

4. Fix a name mismatch with @JsonProperty CORE#

Goal: bind JSON keys "1.symbol", "2.quantity", "3.purchaseDate" to clean fields.

1. Start from a Trade whose fields are symbol/quantity/purchaseDate.
2. Run against the dotted-key JSON, observe nulls/zeros.
3. Add @JsonProperty("…") to each field, re-run.

public class Trade {
    @JsonProperty("1.symbol")       public String symbol;
    @JsonProperty("2.quantity")     public int quantity;
    @JsonProperty("3.purchaseDate") public String purchaseDate;
}

5. Survive an unknown field CORE #

Goal: deserialize JSON that has an extra "4.weather" key without crashing.

1. Use the annotated Trade from Ex 4.
2. Feed it JSON containing "4.weather":"Rainy", observe the exception.
3. Add @JsonIgnoreProperties(ignoreUnknown=true) to the class, re-run.

@JsonIgnoreProperties(ignoreUnknown = true)
public class Trade { /* annotated fields as in Ex 4 */ }

6. Hide a field + handle a duplicate key CHALLENGE#

Goal: never serialize a sensitive field, and observe duplicate-key behavior.

1. Add a password field to a class and mark it @JsonIgnore; serialize and confirm it's absent.
2. Deserialize JSON with a repeated key (e.g. "quantity":100,"quantity":7) and print the value.
3. Explain which one won and why no exception fired.

import com.fasterxml.jackson.annotation.JsonIgnore;

public class User {
    public String username;
    @JsonIgnore public String password;  // skipped both ways
}

ObjectMapper om = new ObjectMapper();

User u = new User(); u.username = "ada"; u.password = "secret";
System.out.println(om.writeValueAsString(u));   // password omitted

String dup = "{\"quantity\":100,\"quantity\":7}";
Trade t = om.readValue(dup, Trade.class);  // Trade w/o @JsonProperty
System.out.println("quantity = " + t.quantity);

Self-check rubric, how to know you've got it#

• I can state serialization vs deserialization in one sentence each, with direction.
• I can write valid JSON from memory and name 3 things JSON forbids.
• I can serialize and deserialize a POJO with ObjectMapper without looking it up.
• Given a null field after deserialization, my first instinct is "key/field name mismatch."
• I know which annotation fixes mismatched keys (@JsonProperty) vs unknown keys (@JsonIgnoreProperties) vs hiding a field (@JsonIgnore).
• I can explain why List<Trade> needs TypeReference.
• I can predict duplicate-key and type-coercion behavior.

★ Key Takeaways#

• Memory is temporary. Collections die with the program; persistence = writing to file/DB/network.
• Cross-language sharing needs a neutral text format, which on the internet is overwhelmingly JSON.
• Serialization = object → format; deserialization = format → object. Same idea as encode/decode.
• JSON = objects {}, arrays [], strings, numbers, booleans, null. Strict punctuation, no trailing commas/comments/single quotes.
• Jackson's ObjectMapper does both directions: writeValue(AsString) out, readValue in. Mapping is by field name.
• POJOs are the schema. Need a no-arg constructor for deserialization; TypeReference for generic collections.
• Annotations bridge the gaps: @JsonProperty (rename), @JsonIgnoreProperties(ignoreUnknown=true) (tolerate extras), @JsonIgnore (hide).
• Edge behavior to memorize: mismatched key → null/default (no error); duplicate key → last wins; unknown field → throws unless ignored; type mismatch → coerced if possible else throws.