I am working on a quiz based project where user visits and gives quiz. I want to list count of users who must have appeared and failed in each of the 5 modules Quizzes. I am getting this properly. Now if a user reappears in any of the 5 quizzes and passes in it he/she must not be included in the earlier mentioned count.
My query for counting number of users appeared and failed in all 5 modules is,
SELECT count(DISTINCT module_id) as module_id from tbl WHERE user_id = $userId AND cleared = 0,
where user_id is user's id and cleared is set to 0 if failed and 1 if passed.
Although the above query counts number of modules that a user has passed. I am using this into a function and then comparing in if statement.
I just want to figure out what changes can I bring in here so that if a user passes in any module, they should not be included in this. I have tried but got no correct query buildup. Any suggestion will be helpful.
Use conditional aggregation:
SELECT user_id
FROM tbl
GROUP BY user_id
HAVING COUNT(DISTINCT module_id) = 5 AND SUM(cleared) = 0;
-- ^^ appeared in all 5 modules ^^ and failed them all
If you want to restrict the above query to just one or a few users, then you may add back the WHERE clause.
Related
I am building a mobile chat web application and I have stuck in a problem. Coming straight to the issue, I have a screen which displays messages list from all users (like WhatsApp). I want to display the last message sent or received between the users in the list (as in the screenshot below). My current query extracts the message from the 1st row for all users right now. That's not what I want.
Little more brief details of what is happening
As you can see in messages table, the fields msg_from and msg_to represents the sender and the receiver respectively. In my data, the messages are transferred between user 1 & 8 and user 1 & 11. For user 1 & 8 the last record fetched should be record 9 which has msg_message Are you there? and similarly, for user 1 & 11 the last record to be fetched would be record 10 which has msg_message Would you like to join?. But currently, for all users the record getting fetched is the 1st record with message How are you?. What changes should my query have to get the desired result? Please have a look at the fiddle below.
Fiddle Here: DB Fiddle
I learned a lot from researching in order to solve this. When grouping, groupBy will take the first row of non-grouped columns (suck as msg_message), so we may order it when joining with the help of a subquery, just like this:
SELECT swp_by, swp_to, msg_from, msg_to, mem_fname, mem_lname, mem_last_activity, msg_message, GREATEST(MAX(msg_time), swipes.swp_date) as msgdate, COUNT(msg_id) as msgcnt FROM swipes
LEFT JOIN
(
SELECT * FROM messages order by msg_time desc -- this is the magic, we use this subquery to order before grouping
)
messages ON
((
messages.msg_from = swipes.swp_by
AND messages.msg_to = swipes.swp_to)
OR (messages.msg_from = swipes.swp_to
AND messages.msg_to = swipes.swp_by
))
solution is in your fiddle: https://www.db-fiddle.com/f/xnh4jiUb8rDLFHpL2gWHrM/5
I think I got expected output
I am struggling to find a good way to approach this code with MYSQL and PHP.
I have a table of users and a table of groups. Each user will only belong to one group, or will belong to group 0, (not in a group).
I have an administration edit group page, which has the JS Picklist Plugin in it.
It lists the users which are in the group, and those that are not.
On submission of the form on this page, I get an array of users which are the only ones which should be in the group after the sql query.
So I need to do two things:
Remove the group id from any users which currently are in the group, but are not in the array returned by the form.
Ensure any users in the array returned have the group id updated to them.
For this example, I have a group ID of 25, after the form has been submitted there are two be two users in this group (44 and 45). No other users should remain in that group. But if the other users are in a different group such as 30, they should not be modified either.
I could have several hundred users, and whilst performance isn't the greatest concern, it just feels a bit hackish to me.
This is what I've failed at so far. For performance reasons we are talking maybe 300 users, so nothing to massive. The two queries do work as expected, just seems very hackish. Must be a better way, here to learn!
update users set user_group = 0 where (user_group = 25 AND uid != 45) AND (user_group = 25 AND uid != 44)
update users set user_group = 25 where ( uid = 44 OR uid = 45 )
Thanks!
UPDATE users
SET user_group = CASE
WHEN uid IN (...ids from post) THEN $group_id
ELSE 0
END
WHERE user_group = $group_id OR uid IN (...ids from post)
Untested and posted from my phone but the concept is there
I have a table with data relating to a user, and two important columns:
refer_count, which is updated when a new entry is made in the table with the referred_by column set to that users user_id, and referred_by which is the user_id of the of the user that referred them.
I want to select the users from the table that have the highest number of referrals after a certain date.
For example:
If there are 3 users, one of which referred the other 2 (lets say users 2 and 3), however user 2 was referred on the 2/12/14, whereas user 3 was referred on the 3/1/15.
If the cutoff is 1/12/14, then user 1 is returned with refer_count set to 2, but if the cutoff is after 2/12/14, then user 1 is returned with refer_count set to 1.
I've been thinking of how to do this, but I can't think of a way that would work. Is there a way?
This is via MySQL.
EDIT: I think I may need to provide for information.
The date registered (register_date) is used as the refer date. I need the refer_count to be updated with the number of users referred after the cutoff, however I need to get the actual user. This is for a 'top referrers' table. I can't figure out why I'm having so much trouble thinking of a way to do this.
SELECT user_id FROM usertable WHERE (referal_date BETWEEN '2014-12-2' AND CURDATE())ORDER BY refer_count DESC;
That's the rough idea.
You should look into normalizing your tables if you're keeping that all in the same table, though. It'd be better to keep referals in a seperate table.
Get the row with the maximum in refer_count with a Date condition for your referal_date such that it's after the certainDate:
SELECT user_id FROM table WHERE refer_count = (SELECT MAX(refer_count) FROM table) AND referal_date>certainDate;
Note that WHERE is before SELECT so it will not get the highest count first, but will filter with the date condition then get the highest count.
Edit: Updated query based on edited question.
I have a web application that stores points in a table, and total points in the user table as below:
User Table
user_id | total_points
Points Table
id | date | user_id | points
Every time a user earns a point, the following steps occur:
1. Enter points value to points table
2. Calculate SUM of the points for that user
3. Update the user table with the new SUM of points (total_points)
The values in the user table might get out of sync with the sum in the points table, and I want to be able to recalculate the SUM of all points for every user once in a while (eg. once a month). I could write a PHP script that could loop through each user in the user table and find the sum for that user and update the total_points, but that would be a lot of SQL queries.
Is there a better(efficient) way of doing what I am trying to do?
Thanks...
A more efficient way to do this would be the following:
User Table
user_id
Points Table
id | date | user_id | points
Total Points View
user_id | total_points
A view is effectively a select statement disguised as a table. The select statement would be: SELECT "user_id", SUM("points") AS "total_points" FROM "Points Table" GROUP BY "user_id". To create a view, execute CREATE VIEW "Total Points View" AS <SELECT STATEMENT> where SELECT STATEMENT is the previous select statement.
Once the view has been created, you can treat it as you would any regular table.
P.S.: I don't know that the quotes are necessary unless your table names actually contain spaces, but it's been a while since I worked with MySQL, so I don't remember it's idiosyncrasies.
You have to user Triggers for this, to make the users total points in sync with the user_points table. Something like:
Create Trigger UpdateUserTotalPoints AFTER INSERT ON points
FOR EACH ROW Begin
UPDATE users u
INNER JOIN
(
SELECT user_id, SUM(points) totalPoints
FROM points
GROUP BY user_id
) p ON u.user_id = p.user_id
SET u.total_points = p.totalPoints;
END;
SQL Fiddle Demo
Note that: As noted by #FireLizzard, if these records in the second table, are frequently updated or delted, you have to have other AFTER UPDATE and AFTER DELETE triggers as well, to keep the two tables in sync. And in this case the solution that #FireLizzard will be better in this case.
If you want it once a month, you can’t deal with just MySQL. You have too « logic » code here, and put too logic in database is not the correct way to go. The trigger of Karan Punamiya could be nice, but it will update the user_table on every insert in points table, and it’s not what you seem to want.
For the fact you want to be able to remove points, just add bsarv new negated rows in points, don’t remove any row (it will break the history trace).
If you really want it periodically, you can run a cron script that does that, or even call your PHP script ;)
The query below selects the 'loves' on an item. (think of it as similar to facebooks 'like' system.
There are two tables in use in this select. A link table (containing itemid, userid, lovetime) and this is joined to a users table in order to retrieve the username/user profile url etc.
$lovequery = "select love.lovetime, love.userid as ID, love_users.display_name, love_users.user_url
from ".$wpdb->prefix."comment_loves love
left join ".$wpdb->prefix."users love_users on love_users.ID=love.userid
where commentid = $itemid
order by love.lovetime desc
limit 4
";
The results are limited to 4 because I simply do not need any more data. The total count is stored separately in the actual item table to reduce queries.
Once the rows are retrieved from this query I iterate through the array, cross referencing against the total 'lovecount' and build a text string formatted like so:
You, John Smith, Joe Bloggs and 4 others love this.
This works fine however it fails if the logged in user (YOU) does not have the most recent 'lovetime'
What I want to do is have the currently logged in use always at the top of the returned results even if his/her 'lovetime' is older than the 4 most recent ones so that the string always begins with 'You' if the logged in user has 'loved' this item.
The logged in user id is available in the script as $userid.
To clarify
if I have the following table (the timestamps are written as simple UK dates for legibility purposes):-
userid commentid lovetime
34 3 02/10/2011
24 3 03/10/2011
13 3 06/10/2011
65 3 14/10/2011
1* 3 10/09/2011
* with userid 1 being the logged in user id
I would only get user id's 34,24,13,65 returned in that order due to ordering by 'lovetime'
What I want is for the results to return ideally 1,34,24,65. if that proves too tricky then getting 5 total rows when the userid exists would be okay also.
I hope this is clear enough, it was rather difficult to articulate.
How would I go about modifying the query to ensure the results are as described.
Many thanks.
You can order result by condition like ORDER BY (ID = "auth_user_id") DESC