What is a Unicode surrogate, and why might it appear in a programming-language string?

A surrogate is a UTF-16 code unit used as part of a pair: a high surrogate and a low surrogate together represent one Unicode code point outside the Basic Multilingual Plane. In JavaScript, strings are sequences of 16-bit UTF-16 code units, and indexes and length count those units. So a supplementary code point can occupy two string indexes while representing one code point. [S1]

A simplified diagram:

One code point outside the BMP:
        ┌─────────────────────┐
        │     one code point  │
        └──────────┬──────────┘
                   │ represented in UTF-16 by
          ┌────────┴────────┐
          │                 │
   high-surrogate     low-surrogate
   code unit 1        code unit 2
          │                 │
        index 0           index 1

JavaScript length: 2 code units
Code-point count:  1

That’s why “string position,” “code point,” and “what looks like one character” can produce different counts. This explanation is specifically about JavaScript’s string model; it should not be assumed to describe every language. [S1]

For a JavaScript string containing one supplementary symbol, inspect both the UTF-16 length and the number of values produced by string iteration:

const s = "\u{1F600}";

console.log(s.length);       // 2: UTF-16 code units
console.log([...s].length);  // 1: values yielded by iteration

The example illustrates the distinction: JavaScript length counts code units, while the string iterator yields code points. [S1] Choose the measurement that matches your task. If you need a JavaScript string index, work in code-unit positions; if you need a count of code points, iterate the string. Don’t treat either result as an automatic count of visible symbols: code-point iteration does not necessarily yield visual grapheme clusters. [S1]

Also be precise about the word “character.” It might mean a UTF-16 code unit, a Unicode code point, or a user-perceived grapheme cluster. State which one you mean before comparing counts or positions.

A lone surrogate is an unpaired UTF-16 code unit. In JavaScript, it does not represent a Unicode character, unlike the high-and-low pair that represents one supplementary code point. [S1]

If you find one, first record the exact string and the operation that exposed it. Check whether your code is viewing individual UTF-16 positions or iterating values, and whether the unit has a matching partner. Don’t infer from a displayed mark or a count alone that the underlying string contains a valid pair. How a particular library, file format, or other programming language handles a lone surrogate depends on that context; the JavaScript description here does not establish those behaviors. [S1]

Use a short note that names the representation and the measurement, then include a minimal example. For instance:

The value is one code point outside the Basic Multilingual Plane, represented in JavaScript as two UTF-16 code units. Therefore `length` and string indexes count 2 here, while string iteration yields 1 code point. This is not necessarily a count of displayed grapheme clusters.

When reporting a real issue, replace “the value” with a safe placeholder or an appropriate test string, and say which operation produced the surprising count. If the code is not JavaScript, verify that language’s string model before using this explanation; JavaScript behavior does not establish how other languages count or index strings. [S1]