The getCollations() method of Intl.Locale instances returns a list of one or more collation types for this locale.

Note: In some versions of some browsers, this method was implemented as an accessor property called collations. However, because it returns a new array on each access, it is now implemented as a method to prevent the situation of locale.collations === locale.collations returning false. Check the browser compatibility table for details.





Return value

An array of strings representing all collation types commonly used for the Locale, sorted in alphabetical order, with the standard and search values always excluded. If the Locale already has a collation, then the returned array contains that single value.

Below is a list of the supported collation types, adapted from the Unicode collation specification.

Supported collation types


Pinyin ordering for Latin, big5 charset ordering for CJK characters (for Chinese)

Warning: The big5han collation type is deprecated, not available in Firefox, Chrome or Edge.


A previous version of the ordering, for compatibility (for Arabic)


Dictionary style ordering (for Sinhala)


Binary code point order

Warning: The direct collation type has been deprecated. Do not use.


The default Unicode collation element table order

Warning: The ducet collation type is not available to the Web. Use the und locale without a collation type specifier instead. und is the collation that is the closest to ducet.


Recommended ordering for emoji characters (for the und locale)


European ordering rules (for the und locale)


Pinyin ordering for Latin, gb2312han charset ordering for CJK characters (for Chinese)

Warning: The gb2312 collation type is deprecated, not available in Firefox, Chrome or Edge.


Phonebook style ordering (for German)


Phonetic ordering (sorting based on pronunciation; for Lingala)


Pinyin ordering for Latin and for CJK characters (for Chinese)


Reformed ordering (formerly for Swedish)

Warning: Do not use explicitly. This is the old name for the default ordering for Swedish whose collation naming used to differ from other languages. Since this was the default, request sv instead of requesting sv-u-co-reformed.

Special collation type for string search

Warning: Do not use as a collation type, since in Intl.Collator, this collation is activated via the "search" value for the usage option. There is currently no API for substring search, so this is currently only good for filtering a list of strings by trying a full-string match of the key against each list item.


Special collation type for Korean initial consonant search

Warning: This collation is not for sorting, even though it is made available through Intl.Collator instantiated with usage "sort" as opposed to usage "search".


Default ordering for each language, except Chinese (and, previously, Swedish)

Warning: Do not use explicitly. In general, it's unnecessary to specify this explicitly and specifying this for Swedish is problematic due to the different meaning for Swedish in the past.


Pinyin ordering for Latin, stroke order for CJK characters (for Chinese)


Traditional style ordering (such as in Spanish)


Radical-stroke ordering for Han characters (for Chinese, Japanese, and Korean). Pinyin ordering for Latin in the case of Chinese.

Note: The unihan collation type is not available in Chrome or Edge.


Pinyin ordering for Latin, zhuyin order for Bopomofo and CJK characters (for Chinese)


Obtaining supported collation types

If the Locale object doesn't have a collation already, getCollations() lists all commonly-used collation types for the given Locale. For examples of explicitly setting a collation, see collation examples.

const locale = new Intl.Locale("zh");
console.log(locale.getCollations()); // ["pinyin", "stroke", "zhuyin", "emoji", "eor"]


Intl Locale Info Proposal
# sec-Intl.Locale.prototype.getCollations

Browser compatibility

BCD tables only load in the browser

See also