Skip to main content

Compairing Objects, Maps, and WeakMaps for lookup tables in JavaScript

I use lookup tables/dictionaries/hashmaps/maps a lot when I’m coding. That’s a data structure that stores key and value pairs that are easily retrievable. Here is an example table for storing and looking up element names by their symbol written in Python that uses a dictionary:

periodic_table["H"] = "Hydrogen"

print(periodic_table["H"])
# Hydrogen

In JavaScript, we have three tools for this pattern—Objects, Maps, and WeakMaps—each with strengths and weaknesses.

Objects

Objects are the standard data type for dictionaries. They are so flexible that they are, in a way, the base prototype for all values in JavaScript. Entries are added and retrieved with dot or bracket notation, and then removed with the delete keyword:

// Declare
let periodicTable = {};

// Set
periodicTable.H = "Hydrogen"; // dot notation
periodicTable["He"] = "Helium"; // bracket notation

// Get
periodicTable.H; // "Hydrogen"
periodicTable["He"]; // "Helium"

// Delete
periodicTable.Li = "Lithium";
delete periodicTable.Li;

Object are not strictly iterable, but you can use for...in to access its keys or the global Object’s methods to get an iterable array: Object.keys(), Object.values(), and Object.entries().

for (let key in periodicTable) {
  console.log(key, periodicTable[key]);
}
// H Hydrogen
// He Helium

Object.keys(periodicTable); // ["H", "He"]
Object.values(periodicTable); // ["Hydrogen", "Helium"]
Object.entries(periodicTable); // [["H", "Hydrogen"], ["He", "Helium"]]

Objects will work as a data structure for every problem you encounter. So in some ways, it is the best of the three options for lookup tables. Some weaknesses of Objects are:

  • Prototype properties can conflict with keys in table
  • Key types limited to strings, numbers, and symbols
  • No guaranteed entry order

Maps

The Map object was added to JavaScript in ES2015 and provides a dedicated native solution for lookup tables. Entries are added, retrieved, and removed with dedicated methods:

// Declare
let periodicTable = new Map();

// Set
periodicTable.set("H", "Hydrogen");
periodicTable.set("He", "Helium");

// Get
periodicTable.get("H"); // "Hydrogen"
periodicTable.get("He"); // "Helium"

// Delete
periodicTable.set("Li", "Lithium");
periodicTable.delete("Li");

Maps are iterable so you can access all of the entries with a for...of loop (similar to Arrays). They also have a Map.forEach() method which provides the same data but (curiously) in the reverse order:

for (let entry of periodicTable) {
  console.log(entry);
}
// ["H", "Hydrogen"]
// ["He", "Helium"]

periodicTable.forEach((value, key) => {
  console.log(key, value);
});
// H Hydrogen
// He Helium

Maps also have Map.keys() and Map.values() methods that return iterators for keys and values, but the use cases for those might be rare. More common are the Map.has() method which returns a boolean of whether a key exists, and Map.size getter which returns the number of key/value pairs.

periodicTable.has("Li"); // false
periodicTable.size; // 2

Some other features that would take more time to illustrate:

  • Preserved entry order
  • More performant than objects for frequent additions and subtractions
  • Support any data structure for keys, including objects and functions (this makes them a good choice for implementing memoization)

It is for these methods/features that Maps really shine as data types of lookup tables. The main weaknesses for Maps are:

  • Not serializable

WeakMaps

WeakMaps were added to JavaScript along with Maps in ES2015. A noticeable difference between the two is that keys in WeakMaps must be objects. Entries are added, retrieved, and deleted with dedicated methods:

// Declare
let periodicTable = new WeakMap();

// Set
let H = { symbol: "H" };
periodicTable.set(H, "Hydrogen");

// Get
periodicTable.get(H); // "Hydrogen"

// Delete
let He = { symbol: "He" };
periodicTable.set(He, "Helium");
periodicTable.delete(He);

Note: There is nothing special about the { symbol: "" } syntax here. I’m just creating an object for each key that makes sense for the periodic-table example. You could use any object or object-like value for the key: {}, [], () => {}

Unlike Maps, WeakMaps are not iterable and there is no way to access its size. The only method provided is WeakMap.has() to receive a boolean of whether that key exists or not:

periodicTable.has("He"); // false

All of these differences are due to a shift in focus. WeakMaps are special in that they weakly reference the objects used as keys. This weak reference means that JavaScript can garbage collect those objects sooner, which will improve the memory performance of your application.

If you have a long-living lookup table, this memory-performance feature may be the perfect solution. Some weaknesses of WeakMaps are:

  • Limited key types
  • No iteration or access to keys/values list
  • Not serializable

Try it out

I created a REPL that includes all of the code examples above for you to explore: replit.com/@SeanMcP/JavaScript-lookup-tables

Conclusion

What is the best tool for implementing lookup tables in JavaScript? The answer will vary depending on your specific use case, but here are some general recommendations:

  1. Map is (probably) the best: Between flexible key types, preserved order, easy iteration, and helpful methods, Maps have all of the features that you need for 90% of lookup tables that you will write in your code. Learning the syntax is a worthwhile investment over basic Objects.

  2. Otherwise use an Object: That said, you could still implement every lookup table with an Object. One scenario where objects shine is for inline lookups:

    function getElementName(symbol) {
  return {
    H: "Hydrogen",
    He: "Helium",
  }[symbol];
}

  3. WeakMap for performance: If you are working in a situation where memory performance is important, then WeakMap’s weakly referenced keys will be helpful. There are other programming problems where WeakMaps shine, but we’re looking specifically for lookup tables.

Now that you know the different tools that JavaScript offers for dictionaries/hash maps/maps/lookup tables, you can choose the best solution in your context.