Question

我有一份调查回复表，其中包括以下选项。非常同意= 5，同意= 4，中立= 3，不同意= 2，非常不同意= 1

该表看起来有点像这样：

ID:  Q1  Q2  Q3  Q4  Qn
1    5   4   5   3   2
2    4   5   2   1   4
3    4   4   3   2   3
4    5   4   3   4   3

我正在使用以下MySQL代码来获取一列并生成3个额外的列，以获得有利，中立和不利的百分比。

Select 
(case when Q1 = 5 or Q1 = 4 then 1 else null end) as Q1Fav,
(case when Q1 = 3 then 1 else null end) as Q1Neu,
(case when Q1 = 2 or Q1 = 1 then 1 else null end) as Q1UnFav
From survey_data

这给了我一栏中Q1的有利反应，另一栏中的中性反应等等。

有没有办法扩展它以使MySQL继续生成这些列，直到它到达Qn？（其中Qn是调查的最后一个问题，也是我列表中的最后一栏）。

提前感谢您的帮助！

Answer 1

- 这应该是你所需要的。

- 您可以进行900万次调查。

- 一个可以有200个问题，而另一个可以有4.无论如何。

create table users
(
    userId int auto_increment primary key,
    fName varchar(100) not null,    -- Stan 
    lName varchar(100) not null     -- Smith
)ENGINE=MyIsam;

create table surveys
(
    surveyId int auto_increment primary key,
    sTitle varchar(255) not null,
    dtCreated date not null -- date created in system
)ENGINE=MyIsam;

-- create table questions
-- (    -- to keep things simple to start, assume questions are not in mysql
    -- that can be for version Beta 0.02
-- );

create table UserTakesSurvey
(   -- A user takes a survey (this is going to be useful)
    -- 1 row for each combo kinda obvious
    userId int not null,
    surveyId int not null,
    dtTaken date not null,

    -- Foreign Key constraints go here:
    -- to Users Table:
    FOREIGN KEY (userId) REFERENCES users(userId),
    -- to Surveys Table:
    FOREIGN KEY (surveyId) REFERENCES surveys(surveyId),
    -- so no row can exist here with a wrong (non-existing) userId or surveyId

    -- addition indexes here:
    primary key (userid,surveyid),
    key (surveyid,userid)
)ENGINE=MyIsam;

create table answers
(   id int auto_increment primary key,
    userId int not null,
    surveyId int not null, -- survey id
    qNum int not null,  -- question number
    answer int not null,    -- 5 to 1

    -- Foreign Key constraints go here:
    -- to Users Table:
    FOREIGN KEY (userId) REFERENCES users(userId),
    -- to Surveys Table:
    FOREIGN KEY (surveyId) REFERENCES surveys(surveyId),
    -- so no row can exist here with a wrong (non-existing) userId or surveyId

    -- Additional indexes go here:
    -- create a composite (combo) key incorporating two columns
    KEY (userId,surveyId),
    KEY (surveyId,userId)
)ENGINE=MyIsam;

insert users (fName,lName) values ('Stan','Smith'),('Sally','Higgins');

-- first survey will have 7 questions hypothetically, 2nd 101 questions
insert surveys (sTitle,dtCreated) values ('Feelings about Milk','2013-01-17');
insert surveys (sTitle,dtCreated) values ('Thoughts on Global Warming',curdate());

-- we now have 2 users, 2 surveys
-- Stan is id=1, Sally is id=2. Ditto for survey numbers starting at 1 and up.
-- in web loop, say PHP or whatever, you are collecting question answers
-- loop thru upon Final Submit and do your inserts into table 'answers' (one row each)

-- Assuming for now surveys themselves for simplicity exist on paper/webpage without questions embedded in database (which is not a problem).

-- But trying to simplify the example. Also note that answers.answer is not null but it can be null (no user answer).

-- One table contains all answers without the need to monkey around with creating columns as you were.
-- That monkeying around would be an answers table for each survey, bad idea.

-- This will perform very fast, and data is easily accessible as if in the format you desire. 
-- It may not be readily obvious how, but then again what are followup questions for :>
-- So that question is something like "How do I get at my data" !

计算有利的百分比

1 个答案: