norm

The norm template normalizes the whole input — folding case and, optionally, stripping accents for the given LOCALE — and returns it as a single token without splitting it into words. Because the entire value becomes one token, it behaves like a normalized keyword: two strings match only if they are equal after normalization.

Use it for exact-match or keyword columns — tags, codes, names, enum-like values — that should still compare case-insensitively or accent-insensitively, rather than for free-text search. For per-word tokenization with the same normalization, use text.

Options

Option	Type	Default	Description
LOCALE	string	—	ICU locale
CASE	string	'none'	Case conversion: 'none', 'lower', 'upper'
ACCENT	boolean	true	Preserve accent marks (false folds them away)

Tokenization

norm always emits exactly one token: the input with case and accents normalized per the options. Spaces and punctuation are kept verbatim — the value is never split. The table below shows how the same input transforms under different option combinations.

Input	CASE	ACCENT	Output token
CAFÉ	'lower'	false	cafe
CAFÉ	'none'	true (default)	CAFÉ (unchanged)
café	'upper'	true	CAFÉ

Because two values collide only when their normalized forms are identical, a norm dictionary with CASE = 'lower' and ACCENT = false makes CAFÉ, Café and cafe all match.

Query

CREATE TEXT SEARCH DICTIONARY norm_dict (    template = 'norm',    locale = 'en_US.UTF-8',    case = 'lower',    accent = false);
SELECT ts_lexize('norm_dict', 'CAFÉ');

Result

 ts_lexize----------- {cafe}

Uppercase normalization, accents preserved

Folding to upper case while keeping accent marks turns café into CAFÉ:

Query

CREATE TEXT SEARCH DICTIONARY norm_upper (    template = 'norm',    locale = 'en_US.UTF-8',    case = 'upper',    accent = true);
SELECT ts_lexize('norm_upper', 'café');

Result

 ts_lexize----------- {CAFÉ}

See also

• text — per-word tokenization with the same case and accent normalization
• keyword — keeps the value as one token without normalizing it
• CREATE TEXT SEARCH DICTIONARY