The COLLATE feature lets you sort STRING values according to language- and country-specific rules, known as collations.

Collated strings are important because different languages have different rules for alphabetic order, especially with respect to accented letters. For example, in German accented letters are sorted with their unaccented counterparts, while in Swedish they are placed at the end of the alphabet. A collation is a set of rules used for ordering and usually corresponds to a language, though some languages have multiple collations with different rules for sorting; for example Portuguese has separate collations for Brazilian and European dialects (pt-BR and pt-PT respectively).


  • You cannot currently use collated strings in indexes or primary keys; doing so causes CockroachDB to crash. If you’re interested in using collated strings in these contexts, you can follow this issue on GitHub to be notified when it’s resolved.

  • Operations on collated strings cannot involve strings with a different collation or strings with no collation. However, it is possible to add or overwrite a collation on the fly.

  • Only use the collation feature when you need to sort strings by a specific collation. We recommend this because every time a collated string is constructed or loaded into memory, CockroachDB computes its collation key, whose size is linear in relationship to the length of the collated string, which requires additional resources.

    However, the COLLATE feature does not require additional disk space because collated strings are stored on disk in the same way as STRING values, i.e.,

Supported Collations

CockroachDB supports the collations provided by Go’s language package. The <collation> argument is the BCP 47 language tag at the end of each line, immediately preceded by //. For example, Afrikaans is supported as the af collation.

SQL Syntax

Collated strings are used as normal strings in SQL, but have a COLLATE clause appended to them.

  • Column syntax: STRING COLLATE <collation>. For example:

    You can also use any of the aliases for STRING.
  • Value syntax: <STRING value> COLLATE <collation>. For example:

    > INSERT INTO foo VALUES ('dog' COLLATE en);


Specify Collation for a Column

You can set a default collation for all values in a STRING column.

For example, you can set a column’s default collation to German (de):

> CREATE TABLE de_names (name STRING COLLATE de);

When inserting values into this column, you must specify the collation for every value:

> INSERT INTO de_names VALUES ('Backhaus' COLLATE de), ('Bär' COLLATE de), ('Baz' COLLATE de);

The sort will now honor the de collation that treats ä as a in alphabetic sorting:

> SELECT * FROM de_names ORDER BY name;
|   name   |
| Backhaus |
| Bär      |
| Baz      |

Order by Non-Default Collation

You can sort a column using a specific collation instead of its default.

For example, you receive different results if you order results by German (de) and Swedish (sv) collations:

> SELECT * FROM de_names ORDER BY name COLLATE sv;
|   name   |
| Backhaus |
| Baz      |
| Bär      |

Ad-Hoc Collation Casting

You can cast any string into a collation on the fly.


However, you cannot compare values with different collations:

pq: unsupported comparison operator: <collatedstring{sv}> < <collatedstring{de}>

You can also use casting to remove collations from values.

> SELECT CAST(name AS STRING) FROM de_names ORDER BY name;

See Also

Data Types

Yes No