行转列，动态列枚举分组

润乾软件

于 2023-01-12 10:02:01 发布

阅读量458

点赞数

CC 4.0 BY-SA版权

分类专栏： JAVA计算文章标签： SQL

本文链接：https://blog.youkuaiyun.com/raqsoft/article/details/128654469

JAVA计算专栏收录该内容

363 篇文章

订阅专栏

该博客讨论了在 MongoDB 中如何实现行转列和动态列枚举分组来统计学生分数。由于 MongoDB 不直接支持这种操作，作者提出了使用 Java、PHP 等编程语言或者 SPL 框架来处理数据，通过示例展示了如何按学校和科目统计各分数段的学生人数。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

【问题】

Hi All,

Thanks for the wonderful support the community gets from this forum.

I am trying to accomplish this in MongoDB. Didn’t think it could get this complicated. thought the problem was interesting to solve.

I am trying to get a count of students by scores for various subjects.

the example below shows 2 subjects, in reality we could run this ad-hoc query for 1 or more subjects. (so cannot “can” and should be real-time)

(for e.g. run the query for a bunch of schools and a bunch of subjects)

the grades are always 1-5 no decimal point

use students

db.studentsummary.insert ({school:‘atl1’, sname : ‘Sean’ , sub1: 4, sub2 :5})

db.studentsummary.insert ({school:‘atl1’, sname : ‘chris’ , sub1: 4, sub2 :3})

db.studentsummary.insert ({school:‘atl1’, sname : ‘becky’ , sub1: 5, sub2 :4})

db.studentsummary.insert ({school:‘atl1’, sname : ‘sam’ , sub1: 5, sub2 :4})

db.studentsummary.insert ({school:‘atl2’, sname : ‘dustin’ , sub1: 2, sub2 :2})

db.studentsummary.insert ({school:‘atl2’, sname : ‘greg’ , sub1: 3, sub2 :4})

db.studentsummary.insert ({school:‘atl2’, sname : ‘peter’ , sub1: 5, sub2 :1})

db.studentsummary.insert ({school:‘atl2’, sname : ‘brad’ , sub1: 2, sub2 :2})

db.studentsummary.insert ({school:‘atl2’, sname : ‘liz’ , sub1: 3, sub2 :null})

Desired Output:(Would like to see how close we could get to the desired output below)

show how many got a 5; how many got a 4 and so on…

I tried quite a bit - trying to group by each subject and run different pipelines based on the subjects chosen for query and let the front end manage the merge and pivot , performance was unacceptable. what not.

help will be very highly appreciated.

【回答】

Mongodb 不直接支持行转列，枚举分组做起来也很麻烦，要把数据读出来再用 Java、Php 等编程语言来实现；还涉及到动态列，实现这样的集合运算也很麻烦。使用 SPL 完成这个需求更容易。以下例子不使用 MongoDB 展现，如果使用 MongoDB 需要用 SPL 的 mongo_open()函数连接数据库，用 mongo_shell() 函数查询原始数据结构

	A	B	C
1	=mongo_open(“mongodb://localhost:27017/local?user=test&password=test”)
2	=mongo_shell(A1,“student.find()”)
3	=[5,4,3,2,1]	>subs=[“sub1”,“sub2”]
4	=A2.group(school)	>newfields=[]
5	for subs	>newfields=newfields\|A3.(A5+“_”+string(~))
6	>result=create(${“school,”+newfields.concat@c()})
7	for A4	>temp=[]
8		for subs	>temp=temp\|A7.align@a(A3,${B8}).(~.len())
9		>temp=[A7.school]\|temp
10		>result.record(temp)
11	>mongo_close(A1)

A1、A2：查询原始数据到 A2 序表

A3、B3：准备动态的分数段和列名

A4：将原始数据按学校分组

A5、B5、A6：生成结果表，结构是 school、sub1_5、sub1_4…sub1_1、sub2_5、sub2_4…sub2_1

A7：按学校循环，准备每次向 result 中插入一条数据（B10）

B8、C8：根据 subs 循环，将 A4 中单个学校各科的学生成绩统计出来并按顺序拼接成序列。统计时将 A7 的某科目（B8）按照 A3 对齐分组

B9：拼接学校名称到上述序列，这样值可以和 result 序表的顺序对应上了

B10：插入数据到结果表

如果非动态列，科目较少，写法简化一点，逻辑与上面的方法大同小异：

	A
1	=mongo_open(“mongodb://localhost:27017/local?user=test&password=test”)
2	=mongo_shell(A1,“student.find()”)
3	=A2.group(school)
4	=[5,4,3,2,1]
5	=A3.new(school,~.align@a(A4,sub1).(~.len()):sub1,~.align@a(A4,sub2).(~.len()):sub2)
6	=A5.new(school,~.sub1(1):sub1\_5,~.sub1(2):sub1\_4,~.sub1(3):sub1\_3,~.sub1(4):sub1\_2,~.sub1(5):sub1\_1,~.sub2(1):sub2\_5,~.sub2(2):sub2\_4,~.sub2(3):sub2\_3,~.sub2(4):sub2\_2,~.sub2(5):sub2\_1)
7	>mongo_close(A1)