Multi-Database Support

Go Audit ships with three dialect implementations. Each generates schema DDL, places parameter placeholders, and chooses a JSON storage strategy appropriate for its database.

PostgreSQL

JSONB for old_values, new_values, request_headers, request_body, response_body, metadata
GIN indexes on old_values and new_values for JSON key lookups
TIMESTAMPTZ with DEFAULT NOW() for created_at
BIGSERIAL primary key
Parameter placeholders: $1, $2, …

MySQL

JSON type for JSON columns
B-tree indexes declared inline in CREATE TABLE
TIMESTAMP with DEFAULT CURRENT_TIMESTAMP for created_at
BIGINT UNSIGNED AUTO_INCREMENT primary key
Table engine: InnoDB, charset utf8mb4
Parameter placeholders: ?

SQLite

TEXT for JSON (stored as JSON-encoded strings)
B-tree indexes
DATETIME with DEFAULT (datetime('now')) for created_at
INTEGER PRIMARY KEY AUTOINCREMENT
Parameter placeholders: ?

Great for local development, embedded deployments, and tests.

Setting the Dialect

The dialect is explicit — there is no auto-detection based on the driver. Pass the constant directly in Config:

audit.New(sqlDB, audit.Config{
    Dialect: audit.PostgreSQL, // or audit.MySQL / audit.SQLite
    // ...
})

Valid values:

audit.PostgreSQL
audit.MySQL
audit.SQLite

Custom Dialects

You can register a custom dialect to support Aurora, AlloyDB, ClickHouse, or any other database. See the Dialect reference for the interface and registration call.

The stored record shape is identical across dialects — same columns, same JSON structure, same index intent. You can migrate an audit database from SQLite to PostgreSQL by copying rows and re-parsing the JSON text into JSONB.