JDBC方言

1. spark 读取Databricks Hive表报错

    val url = "jdbc:databricks:/XXX:443/database_name"

//  使用 Azure Active Directory 令牌进行身份验证
    // p.put("AuthMech", "11");
    //p.put("Auth_Flow", "0");
   //p.put("Auth_AccessToken", "<aad-token>");
   //p.put("PWD", "<persoanl-access-token>");

// 使用个人访问令牌进行身份验证
    val prop = new Properties()
    prop.put("PWD", "XXX")
    prop.put("httpPath", "XXX")
    
    val sql = "XXX"

    spark.read.jdbc(url,sql,prop)

// demo 为Azure Databircks jdbc 连接方式， 其他平台的Databricks 大同小异
// https://learn.microsoft.com/zh-cn/azure/databricks/integrations/jdbc-odbc-bi#authentication

java.sql.SQLDataException: [Databricks][JDBC](10140) Error converting value to BigDecimal.

通过自定义方言解决

    JdbcDialects.registerDialect(new JdbcDialect() {
      override def canHandle(url: String): Boolean = url.startsWith("jdbc:databricks") || url.contains("databricks")
      override def quoteIdentifier(colName: String): String = colName
    })