The SQL always changes, using SPI to provide a static java class maybe not good enough.

As described in #34155
The SQLStatementParserCacheHook serves as an extension point, allowing business applications to capture cache-missed compiled SQL statements and store them in a designated location for later use . This mechanism is designed to enable application developers to use these captured SQL statements for parse cache warm-up before the actual user requests occur ,like right after restarting the application.

Objective: Facilitate application-level SQL parse cache warm-up to improve performance on initial queries.
Implementation: Introduce a hook that business logic can subscribe to, which triggers when there is a cache miss during SQL statement parsing. The hook should provide access to the missed SQL statement for external processing.
Benefit: Developers can preemptively compile and cache SQL statements in their applications, reducing the latency experienced by users on their first interaction with the system.

I do not hope SQL input to be through Java's SPI. ShardingSphere is a database (or database protocol), and all SQL-related content should be provided through standard JDBC.
It is better to use SPI for feature extend only, not user API.

I do not hope SQL input to be through Java's SPI. ShardingSphere is a database (or database protocol), and all SQL-related content should be provided through standard JDBC. It is better to use SPI for feature extend only, not user API.

Thank you for the review.
Therefore, the code presented below does not align with your expectations, is that right?

collect like:

public class CollectSQLStatementParserCacheHook implements SQLStatementParserCacheHook {

    public void finishSuccess(String sql, SQLStatement sqlStatement) {
        asyncSaveSqlToStorage(sql);
    }
}

warmup like:

public class SqlParseWarmupHandler {
    public void warmup() {
        SQLStatementParserEngine sqlParseEngine = SQLStatementParserEngineFactory.getSQLStatementParserEngine(new MySQLDatabaseType(), new CacheOption(), new CacheOption());
        List<String> sqlUsed = loadFromStorage();
        for (String sql: sqlUsed) {
            sqlParseEngine.parse(sql, true);
        }
    }
}