I came across some code that looks like this:
export function foo(arg: string): arg is MyType {
return ...
}
I haven't been able to search for is
in either the docs or google, it's a pretty common word and shows up on basically every page.
What does the keyword do in that context?
See the reference for user-defined type guard functions for more information.
function isString(test: any): test is string{
return typeof test === "string";
}
function example(foo: any){
if(isString(foo)){
console.log("it is a string" + foo);
console.log(foo.length); // string function
}
}
example("hello world");
Using the type predicate test is string
in the above format (instead of just using boolean
for the return type), after isString()
is called, if the function returns true
, TypeScript will narrow the type to string
in any block guarded by a call to the function. The compiler will think that foo
is string
in the below-guarded block (and ONLY in the below-guarded block)
{
console.log("it is a string" + foo);
console.log(foo.length); // string function
}
A type predicate is just used in compile time. The resulting .js
file (runtime) will have no difference because it does not consider the TYPE.
I will illustrate the differences in below four examples.
E.g 1: the above example code will not have a compile error nor a runtime error.
E.g 2: the below example code will have a compile error (as well as a runtime error) because TypeScript has narrowed the type to string
and checked that toExponential
does not belong to string
method.
function example(foo: any){
if(isString(foo)){
console.log("it is a string" + foo);
console.log(foo.length);
console.log(foo.toExponential(2));
}
}
E.g. 3: the below example code does not have a compile error but will have a runtime error because TypeScript will ONLY narrow the type to string
in the block guarded but not after, therefore foo.toExponential
will not create compile error (TypeScript does not think it is a string
type). However, in runtime, string
does not have the toExponential
method, so it will have runtime error.
function example(foo: any){
if(isString(foo)){
console.log("it is a string" + foo);
console.log(foo.length);
}
console.log(foo.toExponential(2));
}
E.g. 4: if we don’t use test is string
(type predicate), TypeScript will not narrow the type in the block guarded and the below example code will not have compile error but it will have runtime error.
function isString(test: any): boolean{
return typeof test === "string";
}
function example(foo: any){
if(isString(foo)){
console.log("it is a string" + foo);
console.log(foo.length);
console.log(foo.toExponential(2));
}
}
The conclusion is that test is string
(type predicate) is used in compile-time to tell the developers the code will have a chance to have a runtime error. For javascript, the developers will not KNOW the error in compile time. This is the advantage of using TypeScript.
Success story sharing
boolean
right?is
keyword is actually casting the type and can catch type errors later in the code. See this example for more info.true
corresponds to a specific type.function isFish(pet: Fish | Bird): pet is Fish { return (pet as Fish).swim !== undefined; }; if (isFish(pet)){fish.swim()}
we could just writeif (pet is Fish){fish.swim()}
.