Converting UTF-8 Strings to Byte Arrays

Unmarshalling JSON requires a byte slice input, while strings are stored as UTF-8 in Go. This article explores the efficient conversion of UTF-8 strings to byte arrays.

Direct Conversion

Go allows converting strings to byte slices, creating a copy of the string's bytes:

s := "some text"
b := []byte(s)

However, for large strings, this approach is inefficient due to the copying.

Using io.Reader

An efficient alternative is using strings.NewReader() to create an io.Reader that reads from the string without copying:

s := `{ "somekey": "somevalue" }`

var result interface{}
err := json.NewDecoder(strings.NewReader(s)).Decode(&result)

This approach avoids copying the string.

Small JSON Texts

For small JSON texts, direct conversion remains a fast option:

s := `{ "somekey": "somevalue" }`

var result interface{}
err := json.Unmarshal([]byte(s), &result)

Note: When reading JSON from an io.Reader (e.g., file or network), pass the io.Reader directly to json.NewDecoder() without intermediate string reads.