If you want to sort an array of objects by multiple keys, you pass the object's
key names in the order you want to sort them to sortby
and get back a function
that you can pass to Array.prototype.sort
.
Goals are:
- work with
Array.prototype.sort
- be typesafe
- be lightweight/small
- be usable in browser and node.js
- do not have dependencies
There are already many
tools
like
this
but I was unable to find any
that is typesafe and Array.prototype.sort
compatible.
npm install --save @pabra/sortby
# or
yarn add @pabra/sortby
Let's assume we have a list of users.
const users = [
{ name: 'Bob', age: 23 },
{ name: 'Alice', age: 23 },
{ name: 'Chloe', age: 42 },
{ name: 'Bob', age: 21 },
];
Now we want to sort this array first by name
and then by age
.
users.sort((left, right) =>
left.name < right.name ? -1
: left.name > right.name ? 1
: left.age < right.age ? -1
: left.age > right.age ? 1
: 0,
);
Just for 2 keys this is already alot of typing and easy to produce/overlook typos - especially if you have mixed ascending and descending order direction. How about this for the same result?
import { by } from '@pabra/sortby';
users.sort(by('name', 'age'));
In the following examples we assume following User
object.
type User = {
name: string;
age: number;
};
Import sortby
import { by } from '@pabra/sortby';
For each key of your object that you want to sort by, you need to pass an
argument to by
function. If you want to sort your array of User
s (User[]
)
ascending by the key "name" and then by "age", you have to pass those key names
as 2 arguments to by
.
by('name', 'age')
That will return a function that you can pass to Array.prototype.sort
with a
signature like this:
(left: User, right: User) => number
To sort descending for a key, you need to pass a tuple of key name and sort
direction. Sort directions are abbreviated with asc
for "ascending" and desc
for "descending". To sort first descending by "name" and then descending by
"age", you write:
by(['name', 'desc'], ['age', 'desc'])
Writing
by('name', 'age')
is just a shorthand for
by(['name', 'asc'], ['age', 'asc'])
If you need to sort User
lists at multiple places the same way, you can
prepare a sortby
function to be reusable.
const userSort = by('name', 'age');
Then you can just use it where needed.
const finalUsers = users
.filter(user => user.age > 23)
.sort(userSort);
sortby
works with generic
types. userSort
from above will have the following (simplified) type signature:
const userSort: (
left: { name: any } & { age: any },
right: { name: any } & { age: any },
) => number
This means, userSort
will work on any object with keys "name" and "age". When
you change the shape of your User
type (eg. "name" => "firstName"), typescript
will mark places where you pass userSort
to sort functions instead of the
place where you defined userSort
.
To show typescript that this filter function is supposed to work on User
objects, you can define it like this:
const userSort = by<User>('name', 'age');
The resulting (simplified) type signature looks like:
const userSort: (
left: User,
right: User,
) => number
If you want to sort by nested objects or need to convert/manipulate a value before comparing, you can pass an extractor function instead of the key name. Let's assume you want to convert all user names to lower case before comping:
by(user => user.name.toLowerCase(), 'age')
For descending order, wrap it in a tuple as above:
by([user => user.name.toLowerCase(), 'desc'], 'age')
Run tests.
npm run test
See them succeed.
PASS __tests__/index.ts
sort by name first
✓ sort by name asc (3 ms)
✓ sort by name asc, age asc (1 ms)
✓ sort by name asc, age desc
✓ sort by name desc, age asc (1 ms)
✓ sort by name desc, age desc
sort explicitly asc
✓ sort by name explicitly asc (1 ms)
✓ sort by name asc, age asc
✓ sort by name asc, age desc (1 ms)
sort by age first
✓ sort by age asc
✓ sort by age asc, name asc
✓ sort by age asc, name desc (1 ms)
✓ sort by age desc, name asc
✓ sort by age desc, name desc
extractor
✓ sort by last reversed name asc, age asc (1 ms)
✓ sort by last reversed name desc, age asc
Test Suites: 1 passed, 1 total
Tests: 15 passed, 15 total
Snapshots: 0 total
Time: 2.371 s
Ran all test suites.