opencc

Overview

UTF8; bundled OpenCC 1.4.2.

Version

Build

You can build the RPM / DEB packages for pg_opencc using pig build:

pig build pkg pg_opencc         # build RPM / DEB packages

Install

You can install pg_opencc directly. First, make sure the PGDG and PIGSTY repositories are added and enabled:

pig repo add pgsql -u          # Add repo and update cache

Install the extension using pig or apt/yum/dnf:

pig install pg_opencc;          # Install for current active PG version

pig ext install -y pg_opencc -v 18  # PG 18
pig ext install -y pg_opencc -v 17  # PG 17
pig ext install -y pg_opencc -v 16  # PG 16
pig ext install -y pg_opencc -v 15  # PG 15
pig ext install -y pg_opencc -v 14  # PG 14

dnf install -y pg_opencc_18       # PG 18
dnf install -y pg_opencc_17       # PG 17
dnf install -y pg_opencc_16       # PG 16
dnf install -y pg_opencc_15       # PG 15
dnf install -y pg_opencc_14       # PG 14

apt install -y postgresql-18-opencc   # PG 18
apt install -y postgresql-17-opencc   # PG 17
apt install -y postgresql-16-opencc   # PG 16
apt install -y postgresql-15-opencc   # PG 15
apt install -y postgresql-14-opencc   # PG 14

Create Extension:

CREATE EXTENSION opencc;

Usage

Sources:

• pg_opencc 0.1.0 README
• Extension control file
• SQL API
• Release notes

opencc, distributed as pg_opencc, converts Simplified Chinese, Traditional Chinese, Taiwan and Hong Kong conventions, and Japanese Shinjitai. Version 0.1.0 embeds OpenCC 1.4.2 and its dictionaries in the shared library; no system OpenCC or external dictionary files are required at runtime.

Enable and convert

Supports PostgreSQL 14–18 and requires a UTF8 database. This trusted extension can be installed by a user with CREATE privilege on the database. Objects use the fixed opencc schema; the extension is not relocatable. Preloading and a server restart are not required.

CREATE EXTENSION opencc;

SELECT opencc.s2twp('鼠标里面的硅二极管坏了，导致光标分辨率降低。');
SELECT opencc.convert('繁體中文', 't2s');
SELECT * FROM opencc.configs;

opencc.convert(input, config) selects a conversion explicitly. Wrappers such as opencc.s2t() and opencc.t2s() provide one function per configuration:

Regional conversion and inspection

opencc.localize(input, target, source) accepts zh-Hans, zh-Hant, zh-CN, zh-TW, zh-HK, and ja. Its default source => 'auto' detects the script, but cannot distinguish Taiwan from Hong Kong vocabulary. Supply the source region when regional phrases matter.

SELECT opencc.localize('滑鼠裡面的矽二極體壞了', 'zh-CN', source => 'zh-TW');
SELECT opencc.detect('中文'), opencc.script_stats('中文');
SELECT * FROM opencc.routes;
SELECT * FROM opencc.ambiguities('大战文丑的时候，他的头发很干燥', 's2t');
SELECT opencc.candidates('发', 's2t');
SELECT * FROM opencc.explain('数据库连接池', 's2twp');

Operation and upgrades

Conversion functions are IMMUTABLE STRICT PARALLEL SAFE within an extension version. They can be used in expression indexes and stored generated columns. PostgreSQL 18 generated columns must explicitly use STORED for user-defined conversion functions.

Dictionary changes in a future extension release can change conversion results. Inspect opencc.version() and review the statements returned by opencc.dependents() when upgrading. Dependency discovery does not find calls hidden inside other user-defined functions.

SELECT * FROM opencc.version();
SELECT * FROM opencc.dependents();

Each backend initializes converters on demand. Optional startup preloading can share their memory across backends; changing shared_preload_libraries requires a restart:

shared_preload_libraries = 'opencc'
opencc.preload = 's2t,t2s,s2twp,tw2sp'

Merge opencc into any existing preload list. opencc.preload also accepts all. Dropping the extension leaves its empty schema behind.