Get each day of the week as columns MySQL and PHP

Question

I am writing a web application in PHP with MySQL.

I have a table called counts and this is how data is stored in that table:

Table: counts

id   counts   location_id   media_id   created_at
--------------------------------------------------
1     50         1             1        2017-03-15
2     30         2             1        2017-03-15
3     80         1             2        2017-03-15
4     20         1             1        2017-03-16
5     100        2             2        2017-03-16

For every unique location_id, media_id and created_at, I store count.

I have another table locations which is like this:

Table: locations

id       name
----------------
1     Location 1
2     Location 2
3     Location 3
4     Location 4
5     Location 5

This is the SQL Query I have at the moment:

select sum(counts.count) as views, locations.name as locations, DAYNAME(counts.created_at) AS weekday from `counts` inner join `locations` on `locations`.`id` = `counts`.`location_id` where `counts`.`created_at` between '2016-12-04' and '2016-12-10' group by `weekday`, `counts`.`location_id`;

This is how the data is displayed:

locations     weekday     views
-----------------------------------
Location 1    Mon          50
Location 1    Tue          30
Location 2    Mon          20
Location 2    Tue          70

I'm creating reports and I would like to run a query such that all the days of the week appear as a column with their corresponding values as the view count for that day of the week. I want something like this:

locations     mon   tue  wed  thu  fri  sat   sun
-------------------------------------------------
Location 1    40    60   51   20   40   20    30
Location 2    80    60   100  24   30   10    5  

Is the above possible in MySQL or I would have to use PHP to achieve that? If so, how do I go about it?

Any help will be appreciated, thanks.

NB: The sample data is not accurate.

Answer 1

It's possible to achieve this result with MySQL, using conditional aggregation.

The trick is to use a conditional test in an expression in the SELECT list to determine whether to return a value of count.

Something like this:

 SELECT l.name                                                  AS `locations`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Mon',c.count,0)) AS `mon`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Tue',c.count,0)) AS `tue`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Wed',c.count,0)) AS `wed`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Thu',c.count,0)) AS `thu`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Fri',c.count,0)) AS `fri`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Sat',c.count,0)) AS `sat`
      , SUM(IF(DATE_FORMAT(c.created_at,'%a')='Sun',c.count,0)) AS `sun`
   FROM `locations` l
   LEFT
   JOIN `counts` c
     ON c.location_id = l.id
    AND c.created_at >= '2016-12-04'
    AND c.created_at  < '2016-12-04' + INTERVAL 7 DAY 
  GROUP BY l.name
  ORDER BY l.name

NOTE:

With the sample data, there are two rows for location_id=1 and created_at='2016-03-15', so this query would return a total of 130 for tue (=50+80), not 50 (as shown in the sample output of the existing query).