增加該功能，純粹是在關(guān)issue的時(shí)候看到了第一個(gè)issue,參看 Is there any plan for JDBC drivers?。

大家討論的時(shí)候，提供了兩個(gè)選擇，一個(gè)是apache calcite, 另外一個(gè)是利用 alibaba 的jdbc 連接池 druid。最后我選擇了使用druid 來(lái)完成。為啥不用第一個(gè)的原因如下：

I have tried calcite to add this feature to es-sql. It's proved to be unsuitable. calcite kind like spark datasource API which provide scanTable,filterTabe,translateTable which means you cannot access original sql(Raw SQL) but some expressions . Also, this means join ,group by action are all done in calcite which really impact performance .

PS: 該功能還只是Demo階段，很多功能還不完備。等穩(wěn)定后會(huì)合并到原項(xiàng)目中。目前經(jīng)過(guò)幾輪PR,

• https://github.com/NLPchina/elasticsearch-sql/pull/269
• https://github.com/NLPchina/elasticsearch-sql/pull/273
• https://github.com/NLPchina/elasticsearch-sql/pull/271

elasticsearch-sql 已經(jīng)支持比較復(fù)雜的SQL語(yǔ)法了。

安裝步驟

• 下載項(xiàng)目 https://github.com/allwefantasy/elasticsearch-sql/tree/jdbc-support
• build jar 包
• 添加jar包到你的項(xiàng)目即可。

代碼示例

public class JDBCTests {
    @Test
    public void testJDBC() throws Exception {
        Properties properties = new Properties();
        properties.put("url", "jdbc:elasticsearch://127.0.0.1:9300/twitter2");
        DruidDataSource dds = (DruidDataSource) ElasticSearchDruidDataSourceFactory.createDataSource(properties);
        dds.setInitialSize(1);
        Connection connection = dds.getConnection();
        PreparedStatement ps = connection.prepareStatement("SELECT split(trim(concat_ws('dd',newtype,num_d)),'dd')[0] as nt from  twitter2");
        ResultSet resultSet = ps.executeQuery();
        while (resultSet.next()) {
            System.out.println(resultSet.getString("nt"));
        }

        ps.close();
        connection.close();
        dds.close();
    }

}