Go: What is the difference between nil and empty slice?

Nil slice — zero value of a slice type, no backing array allocated. Empty slice — has a backing array but length is 0.

A nil slice has len == 0, cap == 0, and its internal pointer is nil. It is the default zero value when you declare a slice without initializing it. You can still use append on a nil slice safely — Go allocates a backing array automatically.

1var s []int           // nil slice
2fmt.Println(s == nil) // true
3fmt.Println(len(s))   // 0
4s = append(s, 1)     // works! now s = [1]

An empty slice has a non-nil pointer to a backing array. Both have len == 0 and behave identically in almost all operations — the difference shows up in JSON serialization and equality checks.

1e := []int{}         // empty slice (non-nil)
2m := make([]int, 0)  // also empty slice
3fmt.Println(e == nil) // false
4fmt.Println(m == nil) // false

JSON behavior:

1data, _ := json.Marshal(nilSlice)    // "null"
2data, _ := json.Marshal(emptySlice)  // "[]"

When to use:

• Nil slice — natural "uninitialized" state; use when you haven't determined the content yet.
• Empty slice — when you want to return an empty JSON array instead of null.

Common mistakes:

• Checking len(s) == 0 without also checking s == nil when nil vs empty matters (e.g., API responses).
• Returning nil when the caller expects [] in JSON — causes null vs empty mismatch.
• Iterating over a nil slice with range — this is safe and executes zero iterations, but beginners often think it panics.