名称值列表对列

Question

我一直在MYSQL中讨论这个问题，因为没有Pivots或Cubes。 我一直在使用案例在审判，它的工作原理部分然而我似乎无法对其进行过滤，只有会员=是，尽管子查询工作，筛选出“否”

CREATE TABLE `nvp` (
`rowid` int(11) NOT NULL AUTO_INCREMENT, 
`groupid` varchar(45) DEFAULT NULL, 
`name` varchar(45) DEFAULT NULL, 
`value` varchar(45) DEFAULT NULL, 
`created` date DEFAULT NULL, 
PRIMARY KEY (`rowid`) 
) ENGINE=InnoDB DEFAULT CHARSET=utf8; 


INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('1', 'Fee', '10', '2014-03-01'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('1', 'Member', 'Yes', '2014-03-01'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('1', 'Type', 'MLB', '2014-03-01'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('2', 'Fee', '10', '2014-03-01'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('2', 'Member', 'Yes', '2014-03-01'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('2', 'Type', 'MLB', '2014-03-01'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('3', 'Fee', '10', '2014-03-02'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('3', 'Member', 'Yes', '2014-03-02'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('3', 'Type', 'MLB', '2014-03-02'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('4', 'Fee', '20', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('4', 'Member', 'Yes', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('4', 'Type', 'NFL', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('5', 'Fee', '50', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('5', 'Member', 'Yes', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('5', 'Type', 'NBA', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('6', 'Fee', '50', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('6', 'Member', 'Yes', '2014-03-03');  
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('6', 'Type', 'NBA', '2014-03-03'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('7', 'Fee', '100', '2014-03-04'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('7', 'Member', 'No', '2014-03-04'); 
INSERT INTO `t2`.`nvp` (`groupid`, `name`, `value`, `created`) VALUES ('7', 'Type', 'NBA', '2014-03-04'); 

所需的输出“透视”表，已过滤，只显示会员以及收取的每笔费用的总和，按日期分组。

Date,  MLB, NFL, NBA 
2014-03-01, 20, 0, 0 
2014-03-02, 10, 0, 0 
3014-03-03, 0, 20, 50 

任何人都有一些指针？ 谢谢。

这是试过SQL：

select tbl2.created, 
    sum(case when tbl2.type='MLB' then fee else null end) as MLB, 
    sum(case when tbl2.type='NBA' then fee else null end) as NBA, 
    sum(case when tbl2.type='NFL' then fee else null end) as NFL 
from (select tbl.* from (select 
    created, 
    sum(case when nvp.name='Fee' then value else null end) as Fee, 
    max(case when nvp.name='Type' then value else null end) as Type, 
    max(case when nvp.name='Member' then value else null end) as member 
from nvp group by groupid) AS tbl where member='Yes') as tbl2 group by created; 

# query to transpose row/columns, uses max where strings are required, this does filter out the No. 
(select tbl.* from (select 
    created, 
    sum(case when nvp.name='Fee' then value else null end) as Fee, 
    max(case when nvp.name='Type' then value else null end) as Type, 
    max(case when nvp.name='Member' then value else null end) as member 
from nvp group by groupid) AS tbl where member='Yes'); 

Answer 1

这是做总总结了基本查询：

select t.created, 
     sum(case when t.type = 'MLB' then fee else null end) as MLB, 
     sum(case when t.type = 'NBA' then fee else null end) as NBA, 
     sum(case when t.type = 'NFL' then fee else null end) as NFL 
from (select created, 
      sum(case when nvp.name = 'Fee' then value+0 else null end) as Fee, 
      max(case when nvp.name = 'Type' then value else null end) as Type, 
      max(case when nvp.name = 'Member' then value else null end) as member 
     from nvp 
     group by groupid 
    ) t 
group by created; 

添加where member = 'Yes'应该工作：

select t.created, 
     sum(case when t.type = 'MLB' then fee else null end) as MLB, 
     sum(case when t.type = 'NBA' then fee else null end) as NBA, 
     sum(case when t.type = 'NFL' then fee else null end) as NFL 
from (select created, 
      sum(case when nvp.name = 'Fee' then value+0 else null end) as Fee, 
      max(case when nvp.name = 'Type' then value else null end) as Type, 
      max(case when nvp.name = 'Member' then value else null end) as member 
     from nvp 
     group by groupid 
    ) t 
where member = 'Yes' 
group by created; 

而且，事实上它根据SQL小提琴（参见here）。这看起来与你所尝试的基本相同。