sortingcollation

Sorting collation refers to the set of rules used to compare and order text strings. Collation is determined by factors such as locale, character set, and the strength or level of comparison. It affects how strings are sorted in lists, how unique constraints are enforced, and how indexes are utilized in databases and search systems. Different environments implement collations with varying levels of linguistic precision and performance.

A common framework for Unicode text is the Unicode Collation Algorithm (UCA). The UCA assigns a sequence of weights to characters that guide comparison. Primary weights determine base characters, secondary weights handle diacritics, tertiary weights manage case distinctions and variants, and quaternary weights may address additional rules such as punctuation. Not all systems expose all levels, but the concept helps explain why sorting can differ across locales and settings.

Locales encode language- and region-specific expectations for ordering. A sorting collation can be binary, comparing raw code units, which is fast but language-agnostic, or linguistic, which follows cultural norms for letter order. Many database systems expose named collations, such as utf8_general_ci or utf8_unicode_ci in MySQL, or rely on system-installed locale data in PostgreSQL. SQLite offers simple collations and supports user-defined ones.

Key considerations when choosing a collation include case sensitivity, accent sensitivity, and Kana or width sensitivity in East Asian languages. Collation may interact with normalization forms (NFC/NFD), though normalization is conceptually distinct from collation. The choice of collation influences sorting results, comparisons, and performance, and should reflect the expected language and user behavior for the data.