mysql select max group by id - mysql

good evening,
i have a table:
A B C
45 1 1
22 2 1
40 3 1
43 1 2
21 2 2
61 3 2
49 4 2
60 5 2
76 1 3
41 2 3
57 3 3
i find max(A) from max(B) group by C. The result should be 60 - max number in A from last row in B from each group (C)
Thank you for your help

If i understand correctly your question you could use an inner join on select max(b):
select max(A)
from my_table m
inner join (
select C,
max(B) act_B
from my_table
group by C
) t on t.act_B = m.B and t.c = m.c

Related

how to make an SQL Join with inequality but just select TOP 1 rows for every mach of the inequality?

I have this tables:
table A:
id
value
1
20
2
15
3
10
table B:
id
value
1
20
2
14
3
10
I want all the pairs where A.value >= than B.value. But for every comparison in the WHERE condition i just want the first match. In the example:
I got this query:
SELECT * FROM A, B
WHERE A.date>=B.date;
A_id
A_value
B_id
B_value
1
20
1
20
1
20
2
14
1
20
3
10
2
15
2
14
2
15
3
10
3
10
3
10
but as i said, i just want the first match of every comparison (asume that a_value and b_value are sorted)
So i want to delete (actually ignore) these values:
A_id
A_value
B_id
B_value
1
20
2
14
1
20
3
10
2
15
3
10
and obtain:
A_id
A_value
B_id
B_value
1
20
1
20
2
15
2
14
3
10
3
10
I think i can achieve the result grouping by A_id and A_value and calculating MAX(B_value) but i dont know if this is efficient.
something like this
SELECT A.id,A.Value,MAX(B_value)
FROM A, B
WHERE A.date>=B.date
GROUP BY A.id,A.value;
So the question is:
Is there a query that can give me the result i need ?
You can use ROW_NUMBER() (available in MySQL 8.x). For example:
select *
from (
select
a.id as a_id, a.value as a_value,
b.id as b_id, b.value as b_value,
row_number() over(partition by a.id order by b.value desc) as rn
from a
join b on a.id = b.id
and a.value >= b.value
) x
where rn = 1

SQL JOIN two tables & second table only gets latest entry & with 2 ids

I have 2 tables.
MARKET TABLE
ID main_key sub_key name created_at
------------------------------------------------------
1 1 1 bottle 1606636000
2 2 1 flask 1606642546
3 2 2 flask 1606650045
4 3 1 can 1606650445
5 3 2 can 1606651546
6 4 1 glass 1606652545
MARKET_UPDATES TABLE
ID main_key sub_key price update
------------------------------------------------------
1 1 1 100 1606665555
2 2 1 120 1606665555
3 2 2 150 1606665555
4 3 1 500 1606665555
5 3 2 550 1606665555
6 4 1 25 1606665555
7 1 1 110 1606665666
8 2 1 135 1606665666
9 2 2 145 1606665666
10 3 1 490 1606665666
11 3 2 440 1606665666
12 4 1 29 1606665666
I've tried this.
SELECT *
FROM market m
JOIN (
SELECT MAX(id) max_id, fk_main_key
FROM market_update
GROUP BY fk_main_key, sub_key
) m_max ON (m_max.fk_main_key = m.main_key)
JOIN market_update mu ON (mu.id = m_max.max_id)
But it multiples it a lot of times and I end up with a lot of the same lines. I think it multiples main_key amount and sub_key amount.
I am trying to JOIN MARKET_UPDATES into MARKET so I can get the latest prices from MARKET_UPDATES, but my issue is that I have 2 id's that I have to check on, main_key & sub_key.
So I have to merge the MARKET table with id(7-12) in MARKET_UPDATES.
here is one way based on your query :
SELECT *
FROM market m
JOIN (
SELECT MAX(id) max_id , fk_main_key, sub_key
FROM market_update
GROUP BY fk_main_key, sub_key
) m_max
ON (m_max.fk_main_key = m.main_key
and m_max.sub_key = m.sub_key)
JOIN market_update mu ON (mu.id = m_max.max_id)
and here is another way using window function RANK() :
SELECT *
FROM market m
JOIN (
SELECT id
, fk_main_key
, sub_key
, rank() over (partition by fk_main_key, sub_key order by id desc ) rnk
FROM market_updat
) mu
ON mu.fk_main_key = m.main_key
and mu.sub_key = m.sub_key
and rnk = 1

Sql join 4 tables count rows

I have 4 tables in similar form. Structures of these tables are like:
id team_id position_id country_id
1 1 1 3
2 1 1 3
3 2 2 3
4 3 3 3
I can count rows of one table with:
SELECT count(position_id) as count1, position_id
FROM players1
where country_id = 3
group by position_id;
Getting result as:
position_id count1
1 54
2 41
3 39
4 32
I want join 4 tables and want to get a result like:
position_id count1 count2 count3 count4
1 54 42 51 61
2 41 40 49 59
3 39 29 44 50
4 32 21 37 47
Can you help me write this sql?
As I have understand you question. Execute this Mysql query.
SELECT
d1.position_id AS Positions_Id,
d1.count1 AS count1,
d2.count1 AS count2,
d3.count1 AS count3,
d4.count1 AS count4
FROM (
SELECT position_id, COUNT(position_id) AS count1
FROM players1
WHERE country_id=3
GROUP BY position_id) AS d1
LEFT JOIN (
SELECT position_id, COUNT(position_id) AS count1
FROM players2
WHERE country_id=3
GROUP BY position_id
) AS d2 ON d2.position_id = d1.position_id
LEFT JOIN (
SELECT position_id, COUNT(position_id) AS count1
FROM players3
WHERE country_id=3
GROUP BY position_id
) AS d3 ON d3.position_id = d1.position_id
LEFT JOIN (
SELECT position_id, COUNT(position_id) AS count1
FROM players4
WHERE country_id=3
GROUP BY position_id
) AS d4 ON d4.position_id = d1.position_id

subtract from values by groups

What is the best way to subtract the lowest value from all values by group?
Something like:
ID Name Value
1 A 10
1 B 40
1 C 100
2 A 20
2 B 80
2 C 90
3 A 4
3 B 7
3 C 8
turn to:
ID Name Value
1 A 0
1 B 30
1 C 90
2 A 0
2 B 60
2 C 70
3 A 0
3 B 3
3 C 4
Use window functions:
select id, name, value - min(value) over (partition by id) as
from t;
If you actually want to update the values, in MySQL, aggregation and join are probably the simplest solution:
update t join
(select id, min(value) as min_value
from t
group by id
) tt
on t.id = tt.id
set value = value - min_value
where min_value <> 0;

Unioning and Joining in 1 Query?

I have the following tables:
TABLE1:
A B C
1 2 3
2 4 6
3 6 9
TABLE2:
A B C
4 8 12
5 10 15
6 12 18
TABLE3:
A D
2 X
4 Y
6 Z
I need one query that gives:
A B C D
1 2 3
2 4 6 X
3 6 9
4 8 12 Y
5 10 15
6 12 18 Z
Is that possible?
I can do it in 2 queries, but the person I'm doing it for wants it in 1.
Thanks!
Try this (example on sqlfiddle):
SELECT x.a, x.b, x.c, d
FROM (
SELECT a, b, c FROM table1
UNION ALL
SELECT a, b, c FROM table2
) x
LEFT JOIN table3 ON ( table3.a = x.a )
Sure:
select v1.*, table3.d
from
(select table1.a, table1.b, table1.c
from table1
union all
select table2.a, table2.b, table2.c
from table2
) v1
left join table3 on v1.a = table3.a