If I have a users table and a question table, where one user can have multiple questions, can I return, for example,
bob | 10 questions
sam | 2 questions
with one query?
Using php with pdo for what it's worth.
users table
userID
name
etc.
questions table
questionID
userID
question
flagged
answered
etc.
I want some fields from the users table and a count of the associated questions on the same row. if it can't be done I'll just use separate queries but I just thought I'd ask for the sake of having tidier code
Something like this:
SELECT u.*, COUNT(*) questions
FROM users u
JOIN QUESTIONS q ON u.userid = q.userid
GROUP BY u.userid
If you only want some columns from the users table, replace u.* with the list of columns you care about.
Assuming you have a table that contains one or many question for each user;
SELECT Count(DISTINCT id) FROM questions
// or
SELECT Count(id) FROM questions GROUP BY user_id
this is what Im talking about:
SELECT u.name, COUNT(*) questions
FROM users u
JOIN questions q ON u.userID = q.userID
GROUP BY u.name
Related
Let's say I have 1000 users for my app. I ask them 100 questions with answers just yes/no and I record those answers in a seperate table.
Now, I want to see people who has given the same answers to at least 20 questions.
What kind of algorithm should I follow in order to do this? What are the relevant keywords for googling?
P.S. I work in a WAMP environment.
Join your answers table to itself, selecting answers which share the same question_id and answer but have a different user_id. Group the rows by both user_ids and use a HAVING clause to exclude those with less than 20 matching answers.
Example where you are looking for users similar to your user with user_id "1":
SELECT DISTINCT a2.user_id FROM answers a
INNER JOIN answers a2
ON a.question_id = a2.question_id
AND a.answer = a2.answer
AND a.user_id != a2.user_id
WHERE a.user_id = 1
GROUP BY a.user_id, a2.user_id
HAVING COUNT(*) >= 20;
Technically you don't need to group by a.user_id in this case but I've left it there in case you want to modify the WHERE clause to return results for more than one a.user_id.
I created two tables, one stores the questions of a quiz, and the other one stores all the answers, that users made.
The first table called "questions" contains the questions:
Field names: id|question
Eg. contents:
1|what's your fav color?
2|what's your fav animal?
The second table named "answers" stores all the answers, that users made:
Fields names: id|questionid|userid|answer
Eg. contents:
1|1|1|Red
1|1|3|Magenta
1|1|4|Green
I'd like to select those questions, that haven't been answered yet by a user.
I store the current user's id in a $_SESSION['id'] session. I tried so many ways, to get these questions, the closest query I've made, was this:
$query = SELECT questions.*, answers.* FROM questions LEFT JOIN answers ON questions.id=answers.questionid WHERE answers.id IS NULL OR answers.userid <> '.$_SESSION['id'];
This won't work, because if there's another userid in the answers table at the same question id, it still selects that row. What could be the problem? Where did I mess up my query?
Thanks in advance for all of your help!
Your user condition is in the wrong place. Since you'll want to try to find a match between the specific user and the question and detect a non match, the user part needs to go inside the ON clause with a null check in the WHERE clause;
SELECT q.*
FROM questions q
LEFT JOIN answers a
ON q.id = a.questionid
AND a.userid = YOUR_USER_ID
WHERE a.id IS NULL
An SQLfiddle to test with.
I've a project that create a q&a website.
I want to show questions by these conditions.
1. Show by the latest question, yeah i know just order by created desc.
2. Show and sort questions by most answers.
3. Show and sort questions by most voted. (like most answers) example.
4. Show questions where unanswered. example
And here is my tables structure in database.
TABLE question
COLUMNS
q_id (primary key)
userid
title
content
created
TABLE answer
COLUMNS
a_id (primary key)
userid
q_id
content
created
TABLE vote
COLUMNS
userid
q_id
created
And each tables it can have a million of rows.
For my 4 questions above I'm trying these SQL(s).
1 Show by the latest question. (solved)
select * from question order by created desc
2 Show and sort by most answers. (seems to slow)
SELECT q.*, COUNT(a.id) as answerCount
FROM question q
LEFT JOIN answer a
ON (q.q_id = a.q_id)
ORDER BY answerCount DESC
3 Show and sort by most voted. (seems to slow).
SELECT q.*, COUNT(v.id) as voteCount
FROM question q
LEFT JOIN vote v
ON (q.q_id = v.q_id)
ORDER BY voteCount DESC
4 Show questions where unanswered. (seems to slow)
SELECT q.*
FROM question q
LEFT JOIN answer a
ON p.q_id = a.q_id
WHERE a.q_id IS NULL ORDER BY q.created DESC
Note: If i use INNER JOIN the rows where count = 0 will not be selected.
As I think, The other websites are commonly have field to count answers and votes already? To make it fast and should i change to this or they have some algorithm which no need to count answer and vote in question table?
TABLE question
COLUMNS
q_id (primary key)
userid
title
content
created
answer_count
votes_count
Help or advice will be truly appreciated.
You can try re-writing your queries, but as MySQL is known for preferring joins over more straight-forward ways, they are not likely to be faster. Here are some queries you can try:
Show and sort by most answers. Use GROUP BY and COUNT(*) to make it plain what you do.
SELECT q.*, COUNT(*) as answerCount
FROM question q
LEFT JOIN answer a ON a.q_id = q.q_id
GROUP BY q.q_id
ORDER BY answerCount DESC;
Show and sort by most answers. Count in a sub-query.
SELECT q.*, (select count(*) from answers a where a.q_id = q.q_id) as answerCount
FROM question q
ORDER BY answerCount DESC;
Show and sort by most answers. Count in a derived table query.
SELECT q.*, a.answerCount
FROM question q
LEFT JOIN (select q_id, count(*) as answerCount from answers group by q_id) a
ON a.q_id = q.q_id
ORDER BY a.answerCount DESC;
Show questions where unanswered. I.e. where no answer EXISTS:
SELECT q.*
FROM question q
WHERE NOT EXISTS (select * from answer a where a.q_id = q.q_id)
ORDER BY q.created DESC;
However, as mentioned, these more straight-forward queries are not necessarily faster. Well, you can give them a try anyhow.
So if re-writing the queries doesn't speed things up, then, yes you can add an answer and a vote count to your question table. This is certainly redundant, but if requirements make such a step necessary, then take it.
Background Information
I have the following table structure:
Participants - id, email, organisation_type_id, job_role_id
Answers - id, participant_id, question_id, answer
Questions - id, question, question_category
There are two other tables, organisation types and job roles which just have an id and a name which are referred to in the above tables.
The tables hold data from when someone fills in a questionnaire (each question having a yes or no answer). Each of the questions also falls into one of two categories (denoted by the question_category field).
When someone completes the questionnaire, it creates a record in the Participants table and for each question it creates a record in the answers table.
The Problem
I want to count the yes answers and the no answers for each question, but based on a particular organisation type (which is held in the participants table).
So for example, if I want to know how many people voted yes and how many votes no who are part of organisation type x, I'd want a query like:
Count all answers to each question, where the participant has a organisation_type_id of x, group by answer
To make it a little more confusing, I also want the row to be included even if there are no answers for that particular question yet. For example, I might have answers for question id x, but none from a participant who is part of organisation_type_id y. I'd want the question returned as a row, but with 0 in the 'answer count' column.
Is my table structure the problem here or is it just a really confusing query? So far I'm using the following query, and then looping over the results with PHP to check if it's part of the organisation or job role that I want, but ideally I'd like to do it all in MySQL.
SELECT * FROM `questions`
JOIN `answers` ON `answers`.`question_id` = `questions`.`id`
JOIN `participants` ON `participants`.`id` = `answers`.`participant_id`
Thanks in advance!
Try this,
SELECT questions.id
,questions.question
,count(answers.id) as totalAns
,Sum(Case When participants.id is not null Then 1 Else 0 End) as WithPart
,Sum(Case When answers.answer = 'Yes' Then 1 Else 0 End) as YesAns
,Sum(Case When answers.answer = 'No' Then 1 Else 0 End) as NoAns
FROM questions
left JOIN answers ON answers.question_id = questions.id
left JOIN participants ON participants.id = answers.participant_id
and participants.organisation_type_id = 'x'
group by questions.id,questions.question
if you want answer for questions and questions without answer you must select from question inner join answer. Below query give you questions with answer yes or no answer:
select p.id, a.answer
from Participants p
inner join Answers a ON p.id = a.participant_id
where a.answer = 'yes' or a.answer = null
You want all questions. So select from questions. You want all answers to the question given by certain participants. So outer join the answers by these criteria. To get both counts (for yes and no) in one record sum them, rather than counting records.
select
q.id,
sum( case when a.answer = 'yes' then 1 ) as yes,
sum( case when a.answer = 'no' then 1 ) as no
from questions q
left outer join answers a on a.question_id = q.question_id and a.participant_id in
(
select p.id
from participants p
where p.organisation_type_id = 'x'
)
group by q.id
order by q.id;
I'm trying to create a simple poll function using php and sql.
I have three tables:
Questions
Which simply contains each question asked
question_id | question_text | created_at
Answers
Which contains each answer for each question
question_id | answer_id | answer_text
Answered Questions
Which records who has voted for each option
question_id | answer_id | user_ip
I'm trying to write a query which will return a single question (the most recent) along with all the possible answers to that question and finally a count of each answer to each question. I know I will have to use a GROUP BY clause and possible LEFT OUTER JOIN, but the exact syntax is eluding me atm.
Any advice would be greatly appreciated. Thanks.
This is very similar to the logic in this article http://www.xaprb.com/blog/2006/12/07/how-to-select-the-firstleastmax-row-per-group-in-sql/.
Essentially you need a subquery which selects the single record / question you are interested in, as well as an outer query to select the information related to that record that you are interested in
(I could post another SQL statement to add to the nice collection that have already been posted, but I thought I'd try and shed some light onto how the other posted queries work)
This query should work on most DBMSs:
select q.question_id, question_text, a.answer_id, a.answer_text, count(user_ip)
from questions q
inner join answers a on (q.question_id = a.question_id)
left join answered_questions aq on (a.question_id = aq.question_id
and a.answer_id = aq.answer_id)
where created_at = (select max(created_at)
from questions
)
group by q.question_id, a.answer_id, q.question_text, a.answer_text
Assuming you're usnig MySQL:
SELECT q.* ,
(
SELECT COUNT(*)
FROM answered_questions aq
WHERE aq.answer_id = a.answer_id
AND aq.question_id = q.question_id
) AS votes
FROM (
SELECT *
FROM question
ORDER BY
created_at DESC
LIMIT 1
) q
LEFT OUTER JOIN
answers a
ON a.question_id = q.question_id
SELECT
questions.question_id,
questions.question_text,
answers.answer_id,
answers.answer_text,
COUNT(answered_questions.user_ip)
FROM
questions,answers,
answered_questions
WHERE
questions.question_id=answers.question_id
AND
questions.question_id=
(SELECT
question_id
FROM questions
ORDER BY questions.created_at
LIMIT 1
)
AND
answered_questions.question_id=questions.question_id
GROUP BY
questions.question_id
should work (although I haven't tested it).