从MySQL documentation
如果FIELDS ESCAPED BY字符为空，则不会转义任何字符，并且NULL将输出为NULL，而不是\N。指定空转义符可能不是一个好主意，特别是如果数据中的字段值包含刚刚给出的列表中的任何字符。

艰难的道路：

SELECT 'CU_CustomerID','CU_UserName','CU_Password','CU_Email','CU_Company','CU_Comment','CU_LastBasket','CON_FirstName','CON_MiddleI','CON_LastName','CON_Address1','CON_Address2','CON_City','CON_State','CON_Province','CON_Country','CON_Zip','CON_Phone1','CON_Phone2'
UNION
SELECT COALESCE(T1.CU_CustomerID,''),COALESCE(T1.CU_UserName,''),COALESCE(T1.CU_Password,''),COALESCE(T1.CU_Email,''),COALESCE(T1.CU_Company,''),COALESCE(T1.CU_Comment,''),COALESCE(T1.CU_ShopperPoints,''),COALESCE(T2.CON_FirstName,''),COALESCE(T2.CON_MiddleI,''),COALESCE(T2.CON_LastName,''),COALESCE(T2.CON_Address1,''),COALESCE(T2.CON_Address2,''),COALESCE(T2.CON_City,''),COALESCE(T2.CON_State,''),COALESCE(T2.CON_Province,''),COALESCE(T2.CON_Country,''),COALESCE(T2.CON_Zip,''),COALESCE(T2.CON_Phone1,''),COALESCE(T2.CON_Phone2,'')
FROM CUSTOMERS AS T1
INNER JOIN CONTACT AS T2 ON T1.CU_CustomerID = T2.CON_CustomerID
WHERE T1.CU_CustomerID BETWEEN 0 AND 1000
INTO OUTFILE  'customers.csv'
FIELDS TERMINATED BY  ','
ENCLOSED BY  '"'

这将默认所有值为空字符串而不是NULL（“\N”）。

以防万一为了避免为每个变量输入“ifnull（var，''）as var”，可以使用以下命令：

SELECT GROUP_CONCAT(CONCAT("IFNULL(",COLUMN_NAME,",'') AS ", COLUMN_NAME))
FROM INFORMATION_SCHEMA.COLUMNS
WHERE TABLE_NAME = '<table name>'
AND TABLE_SCHEMA = '<schema name>'
ORDER BY ORDINAL_POSITION

我也有这个问题，解决如下：

select ifnull(`addr`.`reference`, "") AS `reference`, data2 from ...