Link Search Menu Expand Document

« Return to The Lost Chapters

Table of contents
  1. Record Type
  2. Traversing Records
  3. Traversing NESTED Records
  4. Updating Nested Records

Record Type

The Record type is used to describe a certain kind of object, where all the fields are the same type (usually strings) and all values of those fields are another homogenous type. The concept is basically the same as what other languages might call dictionaries, hash tables, maps, or an associative array. A simple example would be something like a “Record that maps strings to numbers”:

const grades: Record<string, number> = {
    "Ada": 97,
    "Babbage": 68,
    "Captain": 37,
    "Pumpkin": 100
};

console.log(grades);
console.log(grades['Ada']);

const moreGrades = {...grades, "Dr. Bart": 83};

console.log(moreGrades["Dr. Bart"]);

At the end of the day, the actual value stored in grades is just an object, the same as any objects we have seen before. However, from TypeScript’s perspective, the type of fields and values in the object are heavily restricted. You cannot assign non-numbers to the fields of the object stored in grades. However, unlike an Interface, we can add new members as needed.

Ironically, the Record type doesn’t actually meet all the usual requirements of what Wikipedia calls a Record type, since they cannot have heterogenous members (aka two members with different types) and the fields are fixed (i.e., you can add new members to a TypeScript Record). Wikipedia’s definition of Records are more similar to how we use Interfaces.

In fact, we recommend thinking of Records as if they were like Arrays, except you can index using any type (efficiently!) instead of just numbers. Since objects are implemented with Hash Maps, you get O(1) expected runtime performance for indexing, again similar to Arrays. We can traverse Records, add and delete elements, and generally treat them like any other collection.

This is why Records are amazing!

Traversing Records

Going through the elements of a Record ends up being very similar to going through an array. We will still use map, but we must first convert the Record to an Array using one of three possible special built-in functions:

  • Object.keys to iterate through the field names
  • Object.values to iterate through the field values
  • Object.entries to iterate through the field’s names and values at the same time

These functions consume an object (in this case, a Record) and produce an array from its keys and/or values. Let’s look at a simple example:

const GRADES: Record<string, number> = {
    "Ada": 97,
    "Babbage": 68,
    "Captain": 37,
    "Pumpkin": 100
};

function App(): JSX.Element {
  return <ul>
    {Object.keys(GRADES).map((user: string) => (
        <li>User: {user}</li>
    ))}
  </ul>;
}

If GRADES had been an array of strings ["Ada", "Babbage", "Captain", "Pumpkin"] then we could have used .map directly. But since instead we had a Record, we needed to use Object.keys(GRADES) instead.

Now let’s access the values instead of the keys.

const GRADES: Record<string, number> = {
    "Ada": 97,
    "Babbage": 68,
    "Captain": 37,
    "Pumpkin": 100
};

function App(): JSX.Element {
  return <ul>
    {Object.values(GRADES).map((grade: number) => (
        <li>Grade: {grade}</li>
    ))}
  </ul>;
}

Accessing both the keys and the values together gets slightly more complicated. The data returned from Object.entries is an “an array of arrays, where each inner array has length 2, with the first element being each key and the second element being each value.” We use array destructuring to unpack the inner pairs into the parameters.

const GRADES: Record<string, number> = {
    "Ada": 97,
    "Babbage": 68,
    "Captain": 37,
    "Pumpkin": 100
};

function App(): JSX.Element {
  return <ul>
    {Object.entries(GRADES).map(([user, grade]: [string, number]) => (
        <li>{user}: {grade}</li>
    ))}
  </ul>;
}

If you’re having a hard time following the data transformation, it might be helpful to know what the Object.entries(GRADES) expression produces:

[ [ "Ada", 97 ], [ "Babbage", 68 ], [ "Captain", 37 ], [ "Pumpkin", 100 ] ]

An array of arrays, where each inner array is a pair of elements representing each key and value.

Traversing NESTED Records

Just like other collection types, you can nest records inside of records (and arrays and other objects!). This can be disorienting at first, but the rules are consistent with each other. Looking at an example is often helpful.

interface Submission {
    score: number,
    late: boolean
}

// Map assignments to (a Map of students to submissions)
const grades: Record<string, Record<string, Submission>> = {
    "Task 1": {
        "Ada": {
            score: 97,
            late: false
        },
        "Babbage": {
            score: 68,
            late: true
        }
    },
    "Task 2": {
        "Ada": {
            score: 100,
            late: false
        },
        "Babbage": {
            score: 35,
            late: true
        }
    }
};

function App(): JSX.Element {
  return <ul>
    {Object.entries(grades).map(([assignment, submissions]: [string, Record<string, Submission>]) => (
        Object.entries(submissions).map(([user, submission]: [string, Submission]) => (
            <li>{user} scored a {submission.score} on {assignment}</li>
        ))
    ))}
  </ul>;
}

The grades variable is holding three layers of objects: two layers of records and then an interface. The code may be confusing, but if you consider the types they should be helpful in navigating what you have at any given moment.

Updating Nested Records

Beyond traversing, Records are essentially just regular objects. You can use the same operations we learned before on objects to manipulate them. However, there is a trick we can use when creating new objects based off old objects where the field to be edited is dynamic (i.e., a variable instead of a string literal). We wrap the variable inside of square brackets:

interface Submission {
    score: number,
    late: boolean
}

// Use a type variable to avoid having to retype such a long type expression!
type SubmissionRecord = Record<string, Record<string, Submission>>;

// Map assignments to (a Map of students to submissions)
const INITIAL_GRADES: SubmissionRecord = {
    "Task 1": {
        "Ada": {
            score: 97,
            late: false
        },
        "Babbage": {
            score: 68,
            late: true
        }
    },
    "Task 2": {
        "Ada": {
            score: 100,
            late: false
        },
        "Babbage": {
            score: 35,
            late: true
        }
    }
};

function App(): JSX.Element {
    const [grades, setGrades] = useState<SubmissionRecord>(INITIAL_GRADES);

    function editScore(assignment: string, user: string, newScore: number) {
        setGrades({
            // Keep all the other fields the same
            ...grades,
            // But use the assignment parameter. Without the brackets this would be treated as
            //  "assignment", AKA the string literal instead of the variable.
            [assignment]: {
                // Copy all the other fields unchanged...
                ... grades[assignment],
                // But again update the user's grade dynamically
                [user]: {
                    late: grades[assignment][user].late,
                    score: newScore
                }
            }
        })
    }

    return <ul>
    {Object.entries(grades).map(([assignment, submissions]: [string, Record<string, Submission>]) => (
        Object.entries(submissions).map(([user, submission]: [string, Submission]) => (
            <li>
                {user} scored a {submission.score} on {assignment}.
                <Button onClick={()=>editScore(assignment, user, submission.score+1)}>Increase by 1</Button>
            </li>
        ))
    ))}
</ul>;
}

Back to top

Created by Austin Cory Bart, Josh Lyon, Kurt Hammen, Emma Adelmann, Terry Harvey.