Nullable JSON Fields as Go Pointers

Using Pointers for Nullable JSON Fields

JSON supports null as a value, but Go does not have a null type for primitives. The solution is to use pointer types, where nil represents JSON null.

The Problem

{ "name": "Alice", "age": null, "score": 0 }

With a plain int field, both null and 0 unmarshal to 0, making them indistinguishable.

The Solution: Pointers

type Player struct {
    Name  string `json:"name"`
    Age   *int   `json:"age"`
    Score *int   `json:"score"`
}

Now:

• null → Age is nil
• 0 → Score points to an int with value 0
• missing key → Age is nil

Checking Pointer Fields

if player.Age != nil {
    fmt.Printf("Age: %d\n", *player.Age)
} else {
    fmt.Println("Age: not provided")
}

When to Use Pointers

Use pointers when:

• The API sends null for a field and that differs from the zero value
• You need to distinguish "field absent" from "field is zero"
• Building PATCH request bodies where only changed fields are sent

Avoid pointers when:

• The field is always present and non-null
• The zero value is not a valid value in your domain
• Simplicity matters more than null-safety

Combining with omitempty

Age *int `json:"age,omitempty"`

This omits the field when the pointer is nil, includes it (even as 0) when non-nil.

Performance Note

Pointers add a small heap allocation per field. For high-throughput services processing millions of objects, benchmarking may reveal that avoiding unnecessary pointers improves performance.