type Obj = { name: string age: number } const getObjField = (obj: any, key: any) => { return obj[key] } const user: Obj = { name: "Nik", age: 25 } const value1 = getObjField(user, "name") const value2 = getObjField(user, "age") const value3 = getObjField(user, "test") const getObjField = <T, K extends keyof T>(obj: T, key: K) => { return obj[key] }

In the first version, the getObjField function parameters have the type any, which does not provide type checking and can lead to errors, for example, when accessing a non-existent key.

In the second version, a generic with the constraint K extends keyof T is used, which guarantees that the key indeed exists in the obj object. This increases safety and ease of use of the function.

Example:

 type Obj = {
  name: string;
  age: number;
};

const getObjField = <T, K extends keyof T>(obj: T, key: K): T[K] => {
  return obj[key];
};

const user: Obj = { name: "Nik", age: 25 };

const value1 = getObjField(user, "name"); // type string
const value2 = getObjField(user, "age");  // type number
// const value3 = getObjField(user, "test"); // Compilation error, key "test" does not exist in Obj

Thus, the second version of the function ensures type safety and prevents errors at compile time.