Is there a way to make Lodash's orderBy function support accented characters?

Like á, é, ñ, etc. These are moved to the end of the array when the sort is performed.

Is there a way to make Lodash's orderBy function support accented characters?

Like á, é, ñ, etc. These are moved to the end of the array when the sort is performed.

• javascript
• lodash

2 Answers 2

The Problem

It sounds like it doesn't use localeCompare, defaulting instead to the equivalent of using < or >, which pares by UTF-16 code unit numeric values, not locale-aware collation (ordering).

Controlling Comparison Method

You can convert to array (if needed) and then use the native sort with localeCompare. For instance, instead of:

const result = _.orderBy(theArray, ["value"]);

you can do:

const result = theArray.slice().sort((a, b) => a.value.localeCompare(b.value));

or to sort in-place:

theArray.sort((a, b) => a.value.localeCompare(b.value));

localeCompare uses the default collation order for the default locale. Using Intl.Collator, you can have more control over the collation (like case-insensitivity, the handling of accents, the relative position of upper- and lower-case characters, etc.). For instance, if you wanted the default collation for the default locale but with upper-case characters first:

const collator = new Intl.Collator(undefined, {caseFirst: "upper"});
const result = theArray.slice().sort((a, b) => collator.pare(a.value, b.value));

Live Example:

const theArray = [
    {value: "n"},
    {value: "N"},
    {value: "ñ"},
    {value: "á"},
    {value: "A"},
    {value: "a"},
];
const lodashResult = _.orderBy(theArray, ["value"]);
const localeCompareResult = theArray.slice().sort((a, b) => a.value.localeCompare(b.value));
const collator = new Intl.Collator(undefined, {caseFirst: "upper"});
const collatorResult = theArray.slice().sort((a, b) => collator.pare(a.value, b.value));
show("unsorted:", theArray);
show("lodashResult:", lodashResult);
show("localeCompareResult:", localeCompareResult);
show("collatorResult:", collatorResult);

function show(label, array) {
    console.log(label, "[");
    for (const element of array) {
        console.log(`    ${JSON.stringify(element)}`);
    }
    console.log("]");
}

.as-console-wrapper {
    max-height: 100% !important;
}

<script src="https://cdnjs.cloudflare./ajax/libs/lodash.js/4.17.21/lodash.min.js"></script>

Stable vs Unstable Sort

When I first wrote this answer, there was a slight difference between _.orderBy and the native sort: _.orderBy, like _.sortBy, always does a stable sort, whereas at the time of the original answer JavaScript's native sort was not guaranteed to be stable. Since then, though, the JavaScript specification has been modified to require a stable sort (ES2019). So both _.orderBy/_.sortBy and native sort are stable now.

If "stable" vs. "unstable" sort aren't familiar terms: A "stable" sort is one where two elements that are considered equivalent for sorting purposes are guaranteed to remain in the same position relative to each other; in an "unstable" sort, their positions relative to to each other might be swapped (which is allowed because they're "equivalent" for sorting purposes). Consider this array:

const theArray = [
    {value: "x", id: 27},
    {value: "z", id: 14},
    {value: "x", id: 12},
];

If you do an unstable sort that sorts ascending on just value (disregarding id or any other properties the objects might have), there are two valid results:

// Valid result 1: id = 27 remained in front of id = 12
[
    {value: "x", id: 27},
    {value: "x", id: 12},
    {value: "z", id: 14},
]
// Valid result 2: id = 27 was moved after id = 12
[
    {value: "x", id: 12},
    {value: "x", id: 27},
    {value: "z", id: 14},
]

With a stable sort, though, only the first result is valid; the positions of equivalent elements relative to each other remains unchanged.

But again, that distinction no longer matters, since JavaScript's sort is stable now too.

I've solved it by paring a sanitized element.

theArray.sort(function(a, b) {
    return a.toLowerCase().removeAccents().localeCompare(b.toLowerCase().removeAccents());
});

The removeAccents function:

String.prototype.removeAccents = function () {
return this
    .replace(/[áàãâä]/gi,"a")
    .replace(/[éè¨ê]/gi,"e")
    .replace(/[íìïî]/gi,"i")
    .replace(/[óòöôõ]/gi,"o")
    .replace(/[úùüû]/gi, "u")
    .replace(/[ç]/gi, "c")
    .replace(/[ñ]/gi, "n")
    .replace(/[^a-zA-Z0-9]/g," ");
}