How does Go handle errors as values?

Question

Accepted Answer

Go has **no exceptions** for ordinary errors. Instead, errors are **values** returned from functions (typically as the last return value), and callers **explicitly check** them. This makes error handling visible, deliberate, and impossible to silently ignore — a defining Go philosophy.

## The error type and the core pattern

```go
// `error` is a built-in interface: anything with an Error() string method
type error interface {
    Error() string
}

// functions return (result, error) — error is nil on success
func divide(a, b float64) (float64, error) {
    if b == 0 {
        return 0, errors.New("division by zero")   // create an error
    }
    return a / b, nil    // nil = no error
}

// the ubiquitous calling pattern
result, err := divide(10, 0)
if err != nil {              // CHECK the error explicitly
    fmt.Println("error:", err)
    return
}
fmt.Println(result)
```

The `if err != nil { ... }` check appears constantly in Go — every fallible operation is followed by an explicit error check. This verbosity is deliberate: errors are always visible and handled at the call site.

## Creating errors

```go
errors.New("something failed")                       // simple error
fmt.Errorf("failed to process %s: %w", name, err)    // formatted, with WRAPPING (%w)
```

## Wrapping and inspecting errors

```go
// %w WRAPS an underlying error, preserving the chain
if err != nil {
    return fmt.Errorf("reading config: %w", err)   // adds context, keeps original
}

// unwrap and inspect with errors.Is / errors.As
if errors.Is(err, os.ErrNotExist) { ... }          // check for a specific error
var pathErr *os.PathError
if errors.As(err, &pathErr) { ... }                 // extract a specific error type
```

Wrapping (`%w`) adds context as errors propagate up, while `errors.Is`/`errors.As` let you check for or extract specific underlying errors — the modern way to handle error chains.

## Custom error types

```go
type ValidationError struct {
    Field string
}
func (e *ValidationError) Error() string {           // implement the error interface
    return "invalid field: " + e.Field
}
return &ValidationError{Field: "email"}
```

## errors vs panic

```text
error (values) → EXPECTED, recoverable conditions (file missing, bad input) — the norm
panic          → UNEXPECTED, programmer errors / unrecoverable (nil deref, out of bounds)
→ Use errors for normal failure handling; reserve panic for truly exceptional cases.
```

## Why it matters

Go's error-as-value approach is one of its most defining and debated characteristics, fundamentally different from exception-based languages.

By returning errors as values that callers must explicitly check (`if err != nil`), Go makes error handling **visible, local, and impossible to accidentally ignore** — you see exactly where each error is handled, improving reliability at the cost of verbosity.

Understanding this philosophy, the `(result, error)` pattern, error wrapping (`%w`) with `errors.Is`/`errors.As` for adding context and inspecting chains, custom error types, and the distinction between errors (normal, recoverable) and panic (exceptional, programmer errors) is absolutely essential — it permeates every Go program and represents a core mental shift for developers coming from exception-based languages.

It's also among the most common Go interview topics.