INITCAP

Return the input string with the first letter of each word capitalised and every other letter lowercased.

Category: stringReturns: VARCHARDialect: Standard

Syntax

INITCAP(str)

Description

## Overview Returns the input string with the first letter of each word converted to uppercase and all remaining letters converted to lowercase. Word boundaries are defined as the transition from a non-alphanumeric character (whitespace, punctuation, symbols) to an alphanumeric character. INITCAP is typically used for display formatting of names, addresses, and labels that were stored in inconsistent case. INITCAP is a lossy transformation when the source contains legitimately capitalised internal letters (for example 'McDonald' becomes 'Mcdonald', 'iPhone' becomes 'Iphone'). Use it for best-effort presentation, not for preserving names of people, products, or brands exactly. ## Behavior - Returns NULL when the input is NULL. - Returns an empty string when the input is an empty string. - Word boundary is any non-alphanumeric character: spaces, tabs, newlines, hyphens, underscores, slashes, digits adjacent to letters all reset the capitalisation state. - Each word's first character is uppercased using Unicode case mapping; all other characters are lowercased. - Digits and punctuation are preserved in place. A digit does not itself receive a 'case', but a letter immediately after a digit is treated as the start of a word. - Operates on Unicode code points: multi-byte UTF-8 characters are handled correctly and the output is always valid UTF-8. - Idempotent on ASCII words without internal capitals: INITCAP(INITCAP(x)) equals INITCAP(x) for most inputs. ## Compatibility - The definition of a 'word' (alphanumeric run separated by non-alphanumerics) matches the common SQL convention. - Case mapping follows default Unicode rules. Locale-specific title-casing is not applied.

Parameters

Name	Type	Description
`str`		Specifies the input string. Word boundaries are detected at the transition from a non-alphanumeric character to an alphanumeric character.

Examples

-- Basic title case
SELECT INITCAP('hello world');  -- 'Hello World'

-- Mixed case input is normalised
SELECT INITCAP('hELLO wORLD');  -- 'Hello World'

-- Word boundaries from punctuation
SELECT INITCAP('mary-jane watson');  -- 'Mary-Jane Watson'

-- Title-case a column for display
SELECT INITCAP(LOWER(city)) AS city_display
FROM retail.customers.profiles;

-- All uppercase input is forced to title case
SELECT INITCAP('SQL FUNCTION');  -- 'Sql Function'

-- NULL propagates
SELECT INITCAP(CAST(NULL AS VARCHAR));  -- NULL

Pitfalls

INITCAP destroys intentional internal capitalisation. 'McDonald' becomes 'Mcdonald' and 'iPhone' becomes 'Iphone'. Do not apply INITCAP to proper nouns you must preserve exactly.
Apostrophes count as word separators, so O'Brien becomes O'Brien which looks correct, but d'Artagnan becomes D'Artagnan where the capital A may be wrong for the language.
Digits separate words, so an identifier like 'abc123def' becomes 'Abc123Def'. Strip digits first if you do not want mid-token re-capitalisation.
INITCAP is not a spelling or grammar function. It capitalises every word, including articles and prepositions ('of', 'the', 'and'), which is not standard English title case.