1 原始数据
-- ----------------------------
-- Table structure for `t_bm_repeat_purchase`
-- ----------------------------
DROP TABLE IF EXISTS `t_bm_repeat_purchase`;
CREATE TABLE `t_bm_repeat_purchase` (
`months` int(2) DEFAULT NULL COMMENT '月份',
`total` bigint(21) NOT NULL DEFAULT '0' COMMENT '查询月份对应的下一个月后或几个月后的购买用户数',
`seq` bigint(20) DEFAULT NULL COMMENT '序列号',
`next_months` bigint(4) DEFAULT NULL COMMENT 'months 字段对应的第几个月后，1月后，2月后，3月后。。。'
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
-- ----------------------------
-- Records of t_bm_repeat_purchase
-- ----------------------------
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '1170', '2', '2');
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '2144', '2', '3');
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '1012', '2', '4');
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '873', '2', '5');
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '785', '2', '6');
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '1008', '2', '7');
INSERT INTO `t_bm_repeat_purchase` VALUES ('1', '773', '2', '8');
INSERT INTO `t_bm_repeat_purchase` VALUES ('2', '1446', '2', '3');
INSERT INTO `t_bm_repeat_purchase` VALUES ('2', '700', '2', '4');
INSERT INTO `t_bm_repeat_purchase` VALUES ('2', '665', '2', '5');
INSERT INTO `t_bm_repeat_purchase` VALUES ('2', '533', '2', '6');
INSERT INTO `t_bm_repeat_purchase` VALUES ('2', '694', '2', '7');
INSERT INTO `t_bm_repeat_purchase` VALUES ('2', '551', '2', '8');
INSERT INTO `t_bm_repeat_purchase` VALUES ('3', '1530', '2', '4');
INSERT INTO `t_bm_repeat_purchase` VALUES ('3', '1273', '2', '5');
INSERT INTO `t_bm_repeat_purchase` VALUES ('3', '1062', '2', '6');
INSERT INTO `t_bm_repeat_purchase` VALUES ('3', '1367', '2', '7');
INSERT INTO `t_bm_repeat_purchase` VALUES ('3', '1044', '2', '8');
INSERT INTO `t_bm_repeat_purchase` VALUES ('4', '1035', '2', '5');
INSERT INTO `t_bm_repeat_purchase` VALUES ('4', '775', '2', '6');
INSERT INTO `t_bm_repeat_purchase` VALUES ('4', '949', '2', '7');
INSERT INTO `t_bm_repeat_purchase` VALUES ('4', '790', '2', '8');
INSERT INTO `t_bm_repeat_purchase` VALUES ('5', '939', '2', '6');
INSERT INTO `t_bm_repeat_purchase` VALUES ('5', '1304', '2', '7');
INSERT INTO `t_bm_repeat_purchase` VALUES ('5', '1066', '2', '8');
INSERT INTO `t_bm_repeat_purchase` VALUES ('6', '1110', '2', '7');
INSERT INTO `t_bm_repeat_purchase` VALUES ('6', '899', '2', '8');
INSERT INTO `t_bm_repeat_purchase` VALUES ('7', '1589', '2', '8');
要变成
2 用动态查询 ：
SET @EE='';
set @str_tmp='';
SELECT @EE:=CONCAT(@EE,'SUM(IF(next_months=\'',next_months,'\'',',total,null)) AS "',next_months,'",') as aa into @str_tmp
FROM (SELECT DISTINCT next_months FROM t_bm_repeat_purchase) A order by length(aa) desc limit 1;
SET @QQ=CONCAT('SELECT t_bm_repeat_purchase.months,',left(@str_tmp,char_length(@str_tmp)-1),' FROM t_bm_repeat_purchase GROUP BY months ');
PREPARE stmt FROM @QQ;
EXECUTE stmt ;
deallocate prepare stmt;
动态查询结果：这不是最终我们想要的，舍弃这种查询方法，因为前面为空的数据，还要将后面的数据整体向左平移
3 用静态查询
SELECT t.months,
IF(0>num,NULL,SUBSTRING_INDEX(total, ',', 1)) AS '1',
IF(1>num,NULL,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 2),',',-1)) AS '2', -- 这个是算取第1个数
IF(2>num,NULL,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 3) ,',',-1)) AS '3', -- 取第2个数
IF(3>num,NULL,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 4) ,',',-1)) AS '4', -- 取第三个数
IF(4>num,NULL,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 5) ,',',-1)) AS '5',
IF(5>num,NULL,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 6) ,',',-1)) AS '6',
IF(6>num,NULL,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 7) ,',',-1)) AS '7',
IF(7>num,NULL, SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 8) ,',',-1)) AS '8',
IF(8>num,NULL, SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 9) ,',',-1)) AS '9',
IF(9>num,NULL, SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 10) ,',',-1)) AS '10',
IF(10>num,NULL, SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 11) ,',',-1)) AS '11'
FROM
(SELECT a.months,
CHAR_LENGTH(GROUP_CONCAT(a.total))-CHAR_LENGTH(replace(GROUP_CONCAT(a.total),',','')) as num, -- 这个是算每个月有几个逗号
GROUP_CONCAT(a.total ORDER BY a.next_months) AS total
FROM t_bm_repeat_purchase a
GROUP BY a.months ) t;
SELECT a.months,
CHAR_LENGTH(GROUP_CONCAT(a.total))-CHAR_LENGTH(replace(GROUP_CONCAT(a.total),',','')) as num, -- 这个是算每个月有几个逗号
GROUP_CONCAT(a.total ORDER BY a.next_months) AS total
FROM t_bm_repeat_purchase a
GROUP BY a.months 这个语句下查询结果：
对其进行优化
SELECT t.months,
IF(num>=1,SUBSTRING_INDEX(total, ',', 1),NULL) AS '1',
IF(num>=2,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 2),',',-1) ,NULL) AS '2',
IF(num>=3,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 3) ,',',-1),NULL) AS '3',
IF(num>=4,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 4) ,',',-1),NULL) AS '4',
IF(num>=5,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 5) ,',',-1),NULL) AS '5',
IF(num>=6,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 6) ,',',-1),NULL) AS '6',
IF(num>=7,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 7) ,',',-1),NULL) AS '7',
IF(num>=8,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 8) ,',',-1),NULL) AS '8',
IF(num>=9,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 9) ,',',-1),NULL) AS '9',
IF(num>=10,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 10) ,',',-1),NULL) AS '10',
IF(num>=11,SUBSTRING_INDEX(SUBSTRING_INDEX(total, ',', 11) ,',',-1),NULL) AS '11'
FROM
(SELECT a.months,
COUNT(*) as num, -- 这边取每个月分别有多少个数据
GROUP_CONCAT(a.total ORDER BY a.next_months) AS total
FROM t_bm_repeat_purchase a
GROUP BY a.months) t;
4 动态查询和静态查询优缺点
动态的话，我目前没能做到达到最终结果，并且不方便做insert 表 ，但是可以不限多少月，也就是适用于无限数据的
静态的话 对于基数不大的话，比如12个月，6个月这种能较快列举完的比较合适，对于基数大的就不方便，但是方便做insert 表 ，并且静态的我现在可以做到 需求的要求，所以目前采用动态的做法