How Can I Achieve Accent-Insensitive Searches in PostgreSQL?

PostgreSQL's Approach to Accent-Insensitive Searching

Unlike some databases (like Microsoft SQL Server), PostgreSQL doesn't natively support accent-insensitive collations. While PostgreSQL 12 introduced non-deterministic ICU collations offering case and accent insensitivity, these come with performance trade-offs and operational restrictions.

Strategies for Accent-Insensitive Queries in PostgreSQL

Several methods exist to achieve accent-insensitive searching in PostgreSQL:

1. The unaccent Module:

This module provides the unaccent() function, removing accents from strings. This allows queries like:

<code class="language-sql">SELECT * FROM users WHERE unaccent(name) = unaccent('João');</code>

However, unaccent() isn't IMMUTABLE, preventing its use in expression indexes, and it doesn't expand ligatures (e.g., 'Œ').

2. Optimized C Function Wrapper:

To address unaccent()'s limitations, a more efficient solution involves creating an IMMUTABLE C function wrapper:

<code class="language-sql">CREATE OR REPLACE FUNCTION public.f_unaccent(text)
  RETURNS text
  LANGUAGE sql IMMUTABLE PARALLEL SAFE STRICT
RETURN public.immutable_unaccent(regdictionary 'public.unaccent', );</code>

This allows for the creation of expression indexes:

<code class="language-sql">CREATE INDEX users_unaccent_name_idx ON users(public.f_unaccent(name));</code>

Queries then use the wrapped function:

<code class="language-sql">SELECT * FROM users WHERE f_unaccent(name) = f_unaccent('João');</code>

3. Leveraging pg_trgm for Pattern Matching and Ligatures:

For more flexible pattern matching and ligature handling, the pg_trgm module with trigram indexes offers a powerful solution. A trigram GIN index enables case-insensitive searches and similarity detection:

<code class="language-sql">CREATE INDEX users_unaccent_name_trgm_idx ON users
USING gin (f_unaccent(name) gin_trgm_ops);

SELECT * FROM users WHERE f_unaccent(name) LIKE ('%' || f_unaccent('João') || '%');</code>

Note that pg_trgm indexes are more resource-intensive than standard B-tree indexes.

Choosing the optimal approach depends on the specific needs of your application, balancing query performance with index maintenance costs and the requirement for ligature handling.

The above is the detailed content of How Can I Achieve Accent-Insensitive Searches in PostgreSQL?. For more information, please follow other related articles on the PHP Chinese website!