JSON and Serialization

In modern backend development, communicating with other services usually happens via JSON (JavaScript Object Notation). Go's standard library provides a robust encoding/json package that uses reflection to transform Go structs into JSON strings and vice-versa.

1. Struct Tags: The Mapping Layer

Since Go uses PascalCase for exported fields but JSON typically uses camelCase or snake_case, we use Struct Tags to define the mapping between them.

type User struct {
    ID    int    `json:"id"`
    Name  string `json:"full_name"`
    Email string `json:"email,omitempty"` // Hidden if empty
}

• omitempty: This powerful tag tells the encoder to exclude the field from the final JSON if it has its "Zero Value" (e.g., empty string or 0).

2. Encoding (Marshalling)

Marshalling is the process of converting a Go value into its JSON representation.

u := User{ID: 1, Name: "Alice"}
data, err := json.Marshal(u)
fmt.Println(string(data)) // {"id":1,"full_name":"Alice"}

3. Decoding (Unmarshalling)

Unmarshalling is the process of parsing a JSON string into a Go struct.

var u User
err := json.Unmarshal(data, &u) // Note the pointer

• Pointer Requirement: You must pass a pointer to the target variable so that the Unmarshal function can modify its value.

4. Custom Marshalling and Unmarshalling

Sometimes you need to transform data during serialization (e.g., formatting a Date or hiding sensitive fields). You can do this by implementing the Marshaler or Unmarshaler interfaces on your type.

func (n *Name) UnmarshalJSON(data []byte) error {
    // Custom logic here
}

5. Streaming JSON with Decoders

If you are reading from a network connection or a large file, don't use Unmarshal (which loads everything into memory). Use json.Decoder instead.

decoder := json.NewDecoder(response.Body)
err := decoder.Decode(&user)

Interview Pro-Tips: How does json.Marshal work?

If an interviewer asks about performance:

• The Answer: json.Marshal uses Reflection at runtime to inspect the struct and its tags. Because reflection is computationally expensive, serialization can be a bottleneck in high-traffic services.
• The Solution: For performance-critical code, developers sometimes use code-generation tools like easyjson or alternate formats like Protobuf.

Technical Summary

1. Struct Tags: Metadata used to control serialization behavior.
2. Marshal/Unmarshal: The core functions for JSON conversion.
3. Stream: Use Decoders/Encoders when working with io.Reader/Writer.