Go: What is the difference between map and struct?

Map — heterogeneous key-value store with runtime type checking. Struct — homogeneous collection of named fields with compile-time type checking.

Maps are ideal when keys are not known ahead of time (user input, dynamic data). They use hash tables internally, so lookups are O(1) on average. However, maps have overhead per entry and are not safe for concurrent writes.

1config := map[string]string{
2    "log_level": "debug",
3    "port":      "8080",
4}
5// dynamic key — great for user-provided data
6config[os.Getenv("KEY")] = os.Getenv("VALUE")

Structs give you compile-time safety — the compiler knows every field name and type. This catches typos, enables IDE autocompletion, and makes code self-documenting. Structs are also value types, which means no nil-pointer issues.

1type Config struct {
2    LogLevel string `json:"log_level"`
3    Port     string `json:"port"`
4}
5
6cfg := Config{LogLevel: "debug", Port: "8080"}
7cfg.Log_level = "info"  // compile error — typo caught!

When to use:

• Map — dynamic keys, caching, grouping, counting, configuration from unstructured data.
• Struct — domain models, configuration, API payloads, database records.

Common mistakes:

• Using map[string]interface{} to model complex data — use a struct with json tags instead.
• Reading from a map without checking the ok value — silently returns zero value.
• Using maps in concurrent contexts without synchronization — causes race conditions or panics.
• Forgetting that maps cannot be compared with == (except against nil).