本文主要记录在BI和数据分析过程中碰到的生成时间维度的问题，另外也是一个mysql的存储过程基础示例
包含：存储过程基本语法、变量定义、while循环、异常处理

以下存储过程生成了以当前日期为基准前后3650天的日期记录
sql如下：

创建表：

CREATE TABLE `dim_date` (
  `id` int(8) NOT NULL DEFAULT '0',
  `key` date NOT NULL DEFAULT '0000-00-00',
  `year` int(4) NOT NULL,
  `quarter` int(1) NOT NULL,
  `month` int(2) NOT NULL,
  `week` int(1) NOT NULL COMMENT '星期',
  `weekofyear` int(2) NOT NULL COMMENT '一年中的第几周',
  `day` int(2) NOT NULL COMMENT '日',
  `dayofyear` int(3) NOT NULL COMMENT '一年总的第几天',
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

存储过程：

delimiter //
DROP PROCEDURE IF EXISTS getAllDate; 
CREATE PROCEDURE getAllDate()
BEGIN
    DECLARE count int default 0;
    DECLARE startDay DATE DEFAULT date(now());
    DECLARE endDay DATE DEFAULT DATE(NOW());
    -- 定义异常处理方式  http://www.cnblogs.com/cookiehu/p/4994278.html
    DECLARE out_status VARCHAR(200) DEFAULT 'OK';
    DECLARE CONTINUE HANDLER  
    FOR 1062
    SET out_status='Duplicate Entry';
    -- 异常处理方式完毕
    WHILE count<3650 DO
            INSERT INTO `dim_date`(`id`, `key`, `year`, `quarter`, `month`, `week`, `weekofyear`, `day`, `dayofyear`) VALUES (cast(DATE_FORMAT(startDay,'%Y%m%d') as UNSIGNED), startDay, YEAR(startDay), QUARTER(startDay), MONTH(startDay), WEEKDAY(startDay)+1, week(startDay,1), DAY(startDay), DAYOFYEAR(startDay));
            set count = count +1;
            set startDay = DATE_ADD(DATE(now()),INTERVAL count DAY);
            SET endDay = DATE_SUB(DATE(NOW()),INTERVAL count DAY);
            INSERT INTO `dim_date`(`id`, `key`, `year`, `quarter`, `month`, `week`, `weekofyear`, `day`, `dayofyear`) VALUES (cast(DATE_FORMAT(endDay,'%Y%m%d') as UNSIGNED), endDay, YEAR(endDay), QUARTER(endDay), MONTH(endDay), WEEKDAY(endDay)+1, week(endDay,1), DAY(endDay), DAYOFYEAR(endDay));
    END WHILE;
END//
delimiter ;

调用存储过程

-- TRUNCATE table dim_date;
call getAllDate();