Select groups with the three biggest values - mysql

Consider this data:
---+-----------+-------+--+
| id | name | grade | |
+----+-----------+-------+--+
| 13 | Maria | 10 | |
| 18 | Lorenzo | 10 | |
| 2 | Cissa | 10 | |
| 3 | Neto | 9 | |
| 15 | Gabriel | 9 | |
| 10 | Laura | 9 | |
| 12 | Joãozinho | 8 | |
| 16 | Sergio | 8 | |
| 8 | Adriele | 8 | |
| 6 | Jorgito | 8 | |
| 5 | Aline | 8 | |
| 1 | Cintia | 8 | |
| 19 | Fabiana | 7 | |
| 11 | Vinicius | 7 | |
| 9 | Tatiane | 7 | |
| 7 | Chico | 7 | |
| 4 | Marcos | 7 | |
| 14 | Pedro | 6 | |
| 17 | Mauricio | 6 | |
| 20 | Leonardo | 6 | |
+----+-----------+-------+--+
I need the students with the three biggest grades. I think I need to group the data by grade and limit to the top 3 groups.
"SELECT * FROM student GROUP BY grade LIMIT 3" only gives me 3 rows, that's not what I want.
I've tried to use HAVING to filter the groups, but without success. I don't want to set the filter grade>MAX(grade)-2, because theoretically I will not know the grades. But this filter didn't work anyway.
I'am using MySQL. Please help!

You can do this using a join:
select s.*
from student s join
(select grade
from student
group by grade
order by grade desc
limit 3
) g3
on s.grade = g3.grade;
In most databases, you an do this using in:
select s.*
from student s
where s.grade in (select grade
from student
group by grade
order by grade desc
limit 3
);
However, MySQL seems to reject this syntax.

select s1.*
from students s1
join (select distinct grade from students
order by grade desc limit 3) s2 on s1.grade = s2.grade
Alternatively:
select *
from students
where grade >= (select distinct grade from students
order by grade desc limit 2,1)

select s.*
from student s join
(select top 3 grade
from student
group by grade
order by grade desc
) g3
on s.grade = g3.grade;

Related

Selecting COUNT and MAX columns with 2 tables and a bridge table

so what I am trying to do is having 3 tables (pictures, collections, and bridge) with the following columns:
Collections Table:
| id | name |
------------------
| 1 | coll1 |
| 2 | coll2 |
------------------
Pictures Table: (timestamps are unix timestamps)
| id | name | timestamp |
-------------------------
| 5 | Pic5 | 1 |
| 6 | Pic6 | 19 |
| 7 | Pic7 | 3 |
| 8 | Pic8 | 892 |
| 9 | Pic9 | 4 |
-------------------------
Bridge Table:
| id | collection | picture |
-----------------------------
| 1 | 1 | 5 |
| 2 | 1 | 6 |
| 3 | 1 | 7 |
| 4 | 1 | 8 |
| 5 | 2 | 5 |
| 6 | 2 | 9 |
| 7 | 2 | 7 |
-----------------------------
And the result should look like this:
| collection_name | picture_count | newest_picture |
----------------------------------------------------
| coll1 | 4 | 8 |
| coll2 | 3 | 9 |
----------------------------------------------------
newest_picture should always be the picture with the heighest timestamp in that collection and I also want to sort the result by it. picture_count is obviously the count of picture in that collection.
Can this be done in a single statement with table joins and if yes:
how can I do this the best way?
A simple method uses correlated subqueries:
select c.*,
(select count(*)
from bridge b
where b.collection = c.id
) as pic_count,
(select p.id
from bridge b join
pictures p
on b.picture = b.id
where b.collection = c.id
order by p.timestamp desc
limit 1
) as most_recent_picture
from collections c;
A more common approach would use window functions:
select c.id, c.name, count(bp.collection), bp.most_recent_picture
from collections c left join
(select b.*,
first_value(p.id) over (partition by b.collection order by p.timestamp desc) as most_recent_picture
from bridge b join
pictures p
on b.picture = p.id
) bp
on bp.collection = c.id
group by c.id, c.name, bp.most_recent_picture;

COUNT with multiple table query where I've got to display COUNT results

How do I solve this one? None of my queries works and I've tried few.
I've got to return details of two doctors who ordered the most analysis including the number of analysis they have ordered. That last part of the sentence is what really drives me mad.
To work with I've got these two tables:
mysql> select * from DOCTORES;
+-----------+---------------------------+--------------+-----------+
| DNI_DOC | NOMBRE_DOC | ESPECIALIDAD | TELEFONO |
+-----------+---------------------------+--------------+-----------+
| 22888444O | MATEO DÍAZ, RAMÓN | 3 | 659876457 |
| 22909456Y | HERAS PRADO, ANTONIA | 1 | 676234598 |
| 23456398F | GÓMEZ DAVID, ADRIÁN | 1 | 646768454 |
| 25349857H | BURGOS CASA, CANDY | 2 | 659758476 |
| 55776898K | RAMÓN CORONADO,LUIS | 3 | 654364736 |
| 78988484B | CONRADO ALONSO, JOSE | 2 | 645878745 |
| 88647389P | DOMÍNGUEZ GÓMEZ, MANUEL | 1 | 623787343 |
+-----------+---------------------------+--------------+-----------+
and:
mysql> select * from PETICIONES;
+--------+------------+--------+-----------+-----------+
| ID_PET | FECHA_PET | ID_ANA | DNI_PAC | DNI_DOC |
+--------+------------+--------+-----------+-----------+
| 1 | 2008-01-03 | 2 | 71515623A | 23456398F |
| 2 | 2008-05-10 | 2 | 33788976F | 55776898K |
| 3 | 2008-05-08 | 3 | 79876867X | 23456398F |
| 4 | 2008-05-11 | 4 | 44787345H | 55776898K |
| 5 | 2008-05-12 | 2 | 19887234W | 25349857H |
| 6 | 2008-05-05 | 4 | 22897576R | 55776898K |
| 7 | 2008-03-15 | 5 | 44787345H | 88647389P |
| 8 | 2008-03-19 | 1 | 71515623A | 23456398F |
| 9 | 2008-03-26 | 2 | 71515623A | 78988484B |
| 10 | 2008-03-15 | 2 | 19887234W | 88647389P |
| 11 | 2008-03-15 | 3 | 33788976F | 55776898K |
| 12 | 2008-03-26 | 2 | 44787345H | 23456398F |
+--------+------------+--------+-----------+-----------+
I tried this:
select
NOMBRE_DOC
from
DOCTORES
where
DNI_DOC IN (select DNI_DOC
from PETICIONES
group by ID_ANA
order by count(*) desc
limit 2)
group by
DNI_DOC;
and that:
SELECT
DNI_DOC, ID_ANA, COUNT(ID_ANA) AS count
FROM
TIPOS_ANALISIS
WHERE
ID_ANA = (SELECT ID_ANA
FROM
(SELECT
ID_ANA, COUNT(ID_ANA) as AnaCount
FROM
PETICIONES
GROUP BY
by ID_ANA
ORDER BY
AnaCount DESC
LIMIT 1) t1)
GROUP
BY ID_ANA;
and that:
select a.* from DOCTORES a
-> where a.DNI_DOC = ( SELECT b.DNI_DOC from PETICIONES b
-> group by d.ID_ANA
-> order by count(ID_ANA) DESC
-> limit 2 );
you can't even imagine how frustrating it is when you say it's a simple query... Why it isn't simple to me is a mystery (provided that I am really not a dummy).
Expected output would be something like this:
+-----------+---------------------------+--------------+-----------+
| DNI_DOC | NOMBRE_DOC | ID_ANA | count |
+-----------+---------------------------+--------------+-----------+
| 22888444O | MATEO DÍAZ, RAMÓN | 3 | 6 |
+-----------+---------------------------+--------------+-----------+
I believe you are severely overcomplicating this. Join the 2 tables on DNI_DOC field, group by the doctor's id and name (and any other details you may want to include in the select list), and count the number of distinct ID_ANAs.
SELECT d.DNI_DOC, d.NOMBRE_DOC, COUNT(distinct ID_ANA) AS count
FROM DOCTORES d
INNER JOIN PETICIONES t ON d.DNI_DOC=t.DNI_DOC
GROUP BY d.DNI_DOC, d.NOMBRE_DOC
ORDER BY COUNT(distinct ID_ANA) DESC
LIMIT 2
If you need the total number of peticiones per doctor, then remove the distinct from the count.
Here is what I would do :
select d.*, count(p.ID_PET) as TOTAL_PETICIONES from DOCTORES d
inner join PETICIONES p on p.DNI_DOC = d.DNI_DOC
group by d.DNI_DOC
order by count(ID_ANA) DESC
limit 2
This is not tested and write here so please correct me for typos ;)
Using a join will provide a nice output. Grouping by DOCTORES to be able to count the PETICIONES per DOCtORES
EDIT : Something like that. The output won't be good because I have issues to understand the column content.

mysql: order -> limit -> sum... possible?

i am loosing it over the following problem:
i have a table with participants and points. each participant can have up to 11 point entries of which i only want the sum of the top 6.
in this example lets say we want the top 2 of 3
+----+---------------+--------+
| id | participantid | points |
+----+---------------+--------+
| 1 | 1 | 11 |
+----+---------------+--------+
| 2 | 3 | 1 |
+----+---------------+--------+
| 3 | 3 | 4 |
+----+---------------+--------+
| 4 | 2 | 3 |
+----+---------------+--------+
| 5 | 1 | 5 |
+----+---------------+--------+
| 6 | 2 | 10 |
+----+---------------+--------+
| 7 | 2 | 9 |
+----+---------------+--------+
| 8 | 1 | 3 |
+----+---------------+--------+
| 9 | 3 | 4 |
+----+---------------+--------+
as a result i want something like
+---------------+--------+
| participantid | points |
+---------------+--------+
| 2 | 19 |
+---------------+--------+
| 1 | 16 |
+---------------+--------+
| 3 | 8 |
+---------------+--------+
(it should be ordered DESC by the resulting points)
is this at all possible with mysql? in one query?
oh and the resulting participant ids should be resolved into the real names from another 'partcipant' table where
+----+------+
| id | name |
+----+------+
| 1 | what |
+----+------+
| 2 | ev |
+----+------+
| 3 | er |
+----+------+
but that should be doable with a join at some point... i know...
Using one of the answers from ROW_NUMBER() in MySQL for row counts, and then modifying to get the top.
SELECT ParticipantId, SUM(Points)
FROM
(
SELECT a.participantid, a.points, a.id, count(*) as row_number
FROM scores a
JOIN scores b ON a.participantid = b.participantid AND cast(concat(a.points,'.', a.id) as decimal) <= cast(concat(b.points,'.', b.id) as decimal)
GROUP BY a.participantid, a.points, a.id
) C
WHERE row_number IN (1,2)
GROUP BY ParticipantId
Had an issue with ties until I arbitrarily broke them with the id

mysql query using inner joins and subquery

These are the tables:
professor:
+-------+--------+--------+--------+------+
| empid | name | status | salary | age |
+-------+--------+--------+--------+------+
| 1 | Arun | 1 | 2000 | 23 |
| 2 | Benoy | 0 | 3000 | 25 |
| 3 | Chacko | 1 | 1000 | 36 |
| 4 | Divin | 0 | 5000 | 32 |
| 5 | Edwin | 1 | 2500 | 55 |
| 7 | George | 0 | 1500 | 46 |
+-------+--------+--------+--------+------+
works:
+----------+-------+---------+
| courseid | empid | classid |
+----------+-------+---------+
| 1 | 1 | 10 |
| 2 | 2 | 9 |
| 3 | 3 | 8 |
| 4 | 4 | 10 |
| 5 | 5 | 9 |
| 6 | 1 | 9 |
| 2 | 3 | 10 |
| 2 | 1 | 7 |
+----------+-------+---------+
course:
+----------+------------+--------+
| courseid | coursename | points |
+----------+------------+--------+
| 1 | Maths | 100 |
| 2 | Science | 80 |
| 3 | English | 85 |
| 4 | Social | 90 |
| 5 | Malayalam | 99 |
| 6 | Arts | 40 |
+----------+------------+--------+
The question is :
Return list of employees who have taught course Maths or Science but
not both
The query which I wrote is :
select distinct professor.name from professor
inner join works
on professor.empid=works.empid
where works.courseid in
(select courseid from course where coursename ='Maths' or coursename='Science');
The output I received is:
Arun
Benoy
Chacko
Here the employee 'Arun' shouldnt have been displayed as he as taught both Maths and Science.
Please help me out !!
You may use an aggregate COUNT() to check that the total number of DISTINCT courses taught is exactly 1, while still filtering to the two different types of courses. That ensures that only one, never both, is returned.
Because the IN () limits all rows initially returned only to the two desired courses, professors can have a maximum of 2 possible different courses via COUNT(DISTINCT coursename). A HAVING clause then prohibits those with 2 from the final result set.
SELECT
DISTINCT professor.name
FROM
professor
INNER JOIN works ON professor.empid = works.empid
/* Join against course to get the course names */
INNER JOIN course ON works.courseid = course.courseid
WHERE
/* Restrict only to Maths, Science */
course.coursename IN ('Maths', 'Science')
GROUP BY professor.name
/* Only those with exactly one type of course */
HAVING COUNT(DISTINCT course.coursename) = 1
Here is a demonstration: http://sqlfiddle.com/#!2/2e9610/2
You want to use an xor here instead of an or.
select distinct professor.name from professor
inner join works
on professor.empid=works.empid
where works.courseid in
(select courseid from course where coursename ='Maths' xor coursename='Science');

Select rows with alternate ordered field from another table

Given a *students_exam_rooms* table:
+------------+---------+---------+
| student_id | room_id | seat_no |
+------------+---------+---------+
| 1 | 30 | 1001 |
| 2 | 30 | 1002 |
| 3 | 31 | 2001 |
| 4 | 32 | 2002 |
| 5 | 33 | 3001 |
| 6 | 33 | 3002 |
| 7 | 34 | 4001 |
| 8 | 34 | 4002 |
+------------+---------+---------+
And *students_tbl*:
+------------+-------------+------+
| student_id | studen_name | year |
+------------+-------------+------+
| 1 | Eric | 1 |
| 2 | Mustafa | 1 |
| 3 | Michael | 2 |
| 4 | Andy | 2 |
| 5 | Rafael | 3 |
| 6 | Mark | 3 |
| 7 | Jack | 4 |
| 8 | peter | 4 |
+------------+-------------+------+
How can I select from *students_exam_rooms* ordering by *students_tbl.year* but with one after one like this:
+--------------+------+
| student_name | year |
+--------------+------+
| Eric | 1 |
| Michael | 2 |
| Rafael | 3 |
| Jack | 4 |
| Mustafa | 1 |
| Andy | 2 |
| Mark | 3 |
| Peter | 4 |
+--------------+------+
I'm assuming that you want to order by the "occurrence-count" of the year then the year, e.g. all the first-occurrences of all years first, sorted by year, then all second-occurrences of all years also sorted by year, and so on. That would be a perfect case for emulating other RDBMS' analytic / windowing functions:
select *
from (
select
s.studen_name,
s.year,
ser.*,
(
select 1 + count(*)
from students_tbl s2
where s.year = s2.year
and s.student_id > s2.student_id
) rank
from students_tbl s
JOIN students_exam_rooms ser
ON s.student_id = ser.student_id
) i_dont_really_want_to_name_this
order by rank, year
Here it is against a slightly tweaked version of JW's fiddle: http://www.sqlfiddle.com/#!2/27c91/1
Emulating Analytic (AKA Ranking) Functions with MySQL is a good article that gives more background and explanation.
try any of these below:
SELECT a.studen_name, a.year
FROM students_tbl a
INNER JOIN students_exam_rooms b
ON a.student_id = b.student_id
ORDER BY REVERSE(b.seat_no),
a.year
SQLFiddle Demo
by using Modulo
SELECT a.studen_name, a.year
FROM students_tbl a
INNER JOIN students_exam_rooms b
ON a.student_id = b.student_id
ORDER BY CASE WHEN MOD(b.seat_no, 2) <> 0 THEN 0 ELSE 1 END,
a.year
SQLFiddle Demo
Looks to me like you're trying to sort first by seat and then by year. Looking at your students_exam_rooms table, it looks like you started with a simple seat number and prepended year * 1000. So, if we omit the year, it looks like this:
> select * from fixed_students_exam_rooms;
+------------+---------+---------+
| student_id | room_id | seat_no |
+------------+---------+---------+
| 1 | 30 | 1 |
| 2 | 30 | 2 |
| 3 | 31 | 1 |
| 4 | 32 | 2 |
| 5 | 33 | 1 |
| 6 | 33 | 2 |
| 7 | 34 | 1 |
| 8 | 34 | 2 |
+------------+---------+---------+
And if you had that table, your query is simple:
select
student_name, year
from
modified_student_exame_rooms
left join students_tbl using (student_id)
order by
seat_no, year
;
Using the table as you currently have it, it's only slightly more complicated, assuming the "core seat number" doesn't excede 999.
select
student_name, year
from
modified_student_exame_rooms
left join students_tbl using (student_id)
order by
convert(substr(seat_no, 2), unsigned),
year
;