Adventures in Machine Learning

Mastering PostgreSQL Collations and Locale Support

Understanding PostgreSQL Collations and Locale Support

PostgreSQL is an open-source relational database system used for storing and managing large amounts of data. It supports a variety of data types, including text, numeric, temporal, geometric, and array.

One of the key features of PostgreSQL is its support for internationalization and localization, which includes the use of collations and locale support. Collations are used to determine the rules for comparing and sorting text data in a database.

Different collations have different rules for case-sensitivity, accent-sensitivity, and sorting order. PostgreSQL supports a variety of collations, including those based on traditional sorting rules for different languages and regions, as well as those based on modern Unicode rules.

Available Collations on Operating Systems

The available collations on an operating system depend on the system locale settings. The system locale is a set of environment variables that determine the language, region, and character encoding used in system processes and applications.

In POSIX-compliant systems, the system locale is set using the LANG and LC_ environment variables. The LANG variable sets the default language and region, while the LC_ variables set specific subcategories, such as LC_COLLATE for collation rules and LC_TIME for date and time formatting.

In Windows systems, the system locale is set through the Control Panel settings. Windows supports a limited set of collations, mainly those based on the Windows code pages, which are specific to certain regions and languages.

PG_Collation Catalog in PostgreSQL

In PostgreSQL, the pg_collation catalog contains information about the collations available in a database cluster. The catalog includes the collation name, the collation provider (either libc or ICU), the collation version, and the collation options, such as the sensitivity and sorting order.

The catalog can be accessed using the SQL command SELECT from the pg_collation table.

Setting Collation for a Database

When creating a new database in PostgreSQL, you can specify the collation to be used by default for character data in that database. This can be done using the SQL command CREATE DATABASE with the option LC_COLLATE, followed by the name of the collation.

For example, to create a database with the collation rules for German language and region, you can use the following command:

CREATE DATABASE my_database LC_COLLATE ‘de_DE.utf8’;

Note that the collation used in a database must be compatible with the system locale of the operating system.

Setting Collation for a Table

By default, the collation used in a table column is inherited from the database collation. However, you can override it by specifying a different collation for the column when creating the table or altering the table definition.

This can be done using the SQL command ALTER TABLE with the option ALTER COLUMN TYPE and the COLLATE clause, followed by the name of the collation. For example, to set the collation for a column named “name” in a table named “users” to be case-insensitive, you can use the following command:


Changing the Definition of Collation

Once a collation is defined in PostgreSQL, its definition cannot be changed directly. However, you can create a new collation with a different name and options, and then update the database objects that use the old collation to use the new one instead.

This can be done using the SQL command CREATE COLLATION with the option PROVIDER for the new collation, followed by the desired options. For example, to create a new collation named “new_collation” based on the ICU provider with case-insensitive and accent-insensitive sorting rules, you can use the following command:

CREATE COLLATION new_collation (provider = icu, locale = ‘en_US.utf8’, deterministic = true, caseLevel = false, numericOrder = false);

Then, you can update the collation of a table column or an index using the SQL command ALTER TABLE or ALTER INDEX with the option SET, followed by the new collation name.

For example, to change the collation of a column named “name” in a table named “users” to use the new collation instead of the old one, you can use the following command:


Locale Support Initialization in PostgreSQL

Locale support in PostgreSQL refers to the use of language and region settings to format, interpret, and compare data in a database. PostgreSQL uses the operating system locale settings to initialize its locale support, which includes date and time formatting, number formatting, and collation rules.

When initializing a database cluster, you can specify the system locale settings to be used by PostgreSQL.

Language Settings and Subcategories

The system locale in PostgreSQL consists of the language settings and the subcategories. The language settings determine the language and region used for formatting, and the subcategories determine the specific formats and rules used for different types of data.

PostgreSQL supports a variety of language settings, including those based on ISO language codes and those based on regions or dialects. The subcategories in PostgreSQL include LC_COLLATE for collation rules, LC_CTYPE for character classification, LC_MESSAGES for messages and error notifications, LC_MONETARY for currency formats, LC_NUMERIC for numeric formats, and LC_TIME for date and time formats.

Inability to Change Locale after Initialization

Once a database cluster is initialized with a specific locale, it cannot be changed later. Therefore, it is important to choose the appropriate locale settings that match the desired data formats and collation rules before creating a new database cluster.

You can check the current locale settings of a PostgreSQL instance using the SQL command SHOW with the option LC_COLLATE or LC_CTYPE.

Template Database and Error Message

When initializing a new database cluster in PostgreSQL, a template database named “template1” is used as the basis for creating new databases. The template database contains the initial schema and system objects used by all other databases in the cluster.

Therefore, the template database must be initialized with the desired locale settings before creating any other databases. If you try to create a new database with a different locale than the template database after initialization, you will get an error message such as “new encoding (UTF8) is incompatible with the encoding of the template database (SQL_ASCII)”.

This error indicates that the locale of the new database is not compatible with the locale of the template database. Therefore, you need to initialize the template database with the correct locale settings first, or choose a different template database that matches the desired locale settings.

In conclusion, understanding collations and locale support in PostgreSQL is essential for managing data in a multi-language and multi-region environment. By using the available collations and initializing the appropriate locale settings, you can ensure that your database operates correctly for different data types and sorting rules.

Creating a Collation in PostgreSQL

Collations in PostgreSQL are vital for comparing, sorting, and searching text-based data in the database. In case you need a specific collation that is not available out-of-the-box, you may have to create it from scratch.

In this article, we will explain the need for creating a collation and the process for creating it in PostgreSQL.

Need for a Collation

Different regions and languages have varied sorting and comparison rules, which might not always match the built-in collations available in PostgreSQL. In most cases, you’ll need to create a custom collation for specific business requirements.

For instance, you might need to have a case-insensitive collation for sorting and comparing data for an application developed for the English language.

Creating a Collation

You can create your own collation in PostgreSQL using the CREATE COLLATION command. This command is used in conjunction with the pg_collation catalog, which stores the collation information available in the PostgreSQL cluster.

To create a custom collation, you will need to execute a CREATE COLLATION query and provide the specific parameters required to create the custom collation. The CREATE COLLATION command has several parameters that need to be supplied, including:


Name: This is the name given to the new collation. It needs to be unique within the database cluster.

2. Provider: The provider specifies the library responsible for implementing the collation rules.

PostgreSQL provides two providers: libc and ICU. The ICU provider is more advanced and offers additional functionality for handling different language collations.

3. Collate Type: It defines the sorting strategy for the new collation.

There are three types of collation available in PostgreSQL, including deterministic, non-deterministic, and ICU. 4.

Locale: The locale specifies the language and region in which the collation will be used. 5.

CaseLevel: It determines whether the collation should be case-sensitive or case-insensitive. 6.

NumericOrder: Numeric Order specifies whether the collation should consider numeric values or not. After specifying these parameters, you can create your custom collation by executing the CREATE COLLATION command.

The following example shows how you can create a custom case-insensitive collation:

CREATE COLLATION case_insensitive (


LOCALE = ‘en_US.utf8’,

COLLATE = icu,

CASELEVEL = false,




In this sample query, we created a new collation called ‘case_insensitive.’ We specified that the provider be ICU and set the locale to English-United States. We also mentioned that the collate type be ICY, which enables more advanced collation rules.

Then, we defined our required collation options and set CaseLevel to false to mark it as case-insensitive. Finally, we set the Numeric order to false to omit numbers from sorting.

One important thing to note is that once a collation is created, you cannot modify it. If you want to change the collation rules of an existing collation, you will need to create a new one with the correct parameters.

After creating a new collation, it is possible to update the existing columns or tables’ collation to the new collation with an ALTER TABLE or ALTER COLUMN query.

Wrapping Up

Creating custom collations in PostgreSQL can be helpful in situations where you need to implement specific sorting and comparison rules not present in the default collations. By providing the parameters required for your custom collation, you can use existing PostgreSQL infrastructure to generate a new collation that meets your requirements.

Nonetheless, it’s essential to remember that once created, custom collations cannot be modified. Therefore, if a modification is necessary, a new collation must be added, and any pre-existing entities using the old collation must also be updated with the new one.

Creating a custom collation in PostgreSQL is essential when built-in collations don’t meet specific sorting and comparison rules needed within your database. Collations establish specific rules regarding data sorting to assist in the correct selection of results and data comparison as needed.

By defining specific parameters within the pg_collation catalog and utilizing CREATE COLLATION command, PostgreSQL infrastructure can generate a new collation that fits your database’s needs. Remember that once a custom collation is created, it can’t be modified, and if modifications are necessary, a new collation must replace the existing one.

By following these steps, you can ensure your data is accurately sorted and compared according to your specific criteria, ensuring your PostgreSQL-based infrastructure runs more efficiently.

Popular Posts