I have a MySQL query that selects data from multiple tables and then orders the results based on some arbitrary criteria as below:
SELECT [columns] FROM (
SELECT *, COUNT(*) as `matches`
FROM [table1]
JOIN [table2] USING (id)
JOIN [table3] USING (id)
WHERE [criteria]
GROUP BY `id`
ORDER BY `matches` DESC
) AS `grouped`
ORDER BY (
(CASE WHEN [1st rank criteria] THEN 3 ELSE 0 END) +
(CASE WHEN [2nd rank criteria] THEN 2 ELSE 0 END) +
(CASE WHEN [3rd tank criteria] THEN 1 ELSE 0 END)
) DESC
LIMIT 100
This works fine, but my question is: can I have the ranking score displayed as a column? I've looked at trying to use variables, but I'm quite new to SQL so all this is a bit beyond me.
Sorry if this is an obvious question, but many thanks in advance for your time and assistance.
Try this:
SELECT [columns],
(
(CASE WHEN [1st rank criteria] THEN 3 ELSE 0 END) +
(CASE WHEN [2nd rank criteria] THEN 2 ELSE 0 END) +
(CASE WHEN [3rd tank criteria] THEN 1 ELSE 0 END)
) AS MyRank
FROM (
SELECT *, COUNT(*) as `matches`
FROM [table1]
JOIN [table2] USING (id)
JOIN [table3] USING (id)
WHERE [criteria]
GROUP BY `id`
ORDER BY `matches` DESC
) AS `grouped`
ORDER BY MyRank DESC
LIMIT 100;
put it in the select
((CASE WHEN [1st rank criteria] THEN 3 ELSE 0 END) +
(CASE WHEN [2nd rank criteria] THEN 2 ELSE 0 END) +
(CASE WHEN [3rd tank criteria] THEN 1 ELSE 0 END)) as ranking
Related
SELECT SUM(T1.approve_con) AS compl_rec,
SUM(T1.payout) AS payout,
SUM(T1.reversal) AS reversal,
T1.app_id,
T1.end_user,
T1.created
FROM (
(SELECT app_id,end_user,created,
approve_conversions AS `approve_con`,
'0' AS reversal,
payout AS payout
FROM app_end_users
WHERE publisher_id=1625
AND app_id=57)
UNION ALL
(SELECT opinongold_survey_response.app_id,
opinongold_survey_response.user_id AS end_user,
opinongold_survey_response.movement AS created,
(CASE WHEN opinongold_survey_response.survey_status= 'Completed' THEN 1 ELSE 0 END) AS `approve_con`,
(CASE WHEN opinongold_survey_response.survey_status= 'Rejected' THEN 1 ELSE 0 END) AS reversal,
(CASE WHEN opinongold_survey_response.survey_status= 'Completed' THEN payout ELSE 0 END) AS payout
FROM opinongold_survey_response
WHERE pub_id=1625
AND app_id=57)
) AS T1
GROUP BY T1.app_id,T1.end_user
ORDER BY `approve_con` DESC
LIMIT 0, 10
above is my mysql query order by is not working on approve_con, payout, reversal column
please let me where i am doing wrong in this query.
I've got a simple query which brings up wins, draws and losses in a head-to-head matches table.
SELECT
SUM(CASE WHEN score_w > score_m THEN 1 ELSE 0 END) AS wins_w,
SUM(CASE WHEN score_m > score_w THEN 1 ELSE 0 END) AS wins_m,
SUM(CASE WHEN score_w = score_m THEN 1 ELSE 0 END) AS draws
FROM 6dos7me3xn8
All is well. I get a single row, with the data I need as columns.
Now I want to also pull out a group concat'd list of the most recent three match dates. I tried:
SELECT
SUM(CASE WHEN mn.score_w, 0) > mn.score_m THEN 1 ELSE 0 END) AS wins_w,
SUM(CASE WHEN mn.score_m > mn.score_w THEN 1 ELSE 0 END) AS wins_m,
SUM(CASE WHEN mn.score_w = mn.score_m THEN 1 ELSE 0 END) AS draws,
GROUP_CONCAT(jn.date) AS recent
FROM 6dos7me3xn8 mn
JOIN (SELECT date FROM 6dos7me3xn8 ORDER BY date DESC LIMIT 3) jn
...but the LIMIT is having no effect, seemingly - I get all of the dates group concat'd, not just 3.
I also tried removing the JOIN and replacing the GROUP_CONCAT with
GROUP_CONCAT((SELECT date FROM 6dos7me3xn8 ORDER BY date DESC LIMIT 3)) AS recent
...but that errors with 'Subquery returns more than 1 row.'
I'm sure it's something simple, but what am I doing wrong?
If you are running MySQL 8.0, you can do this with window functions:
select
sum(score_w > score_m) as wins_w,
sum(score_m > score_w) as wins_m,
sum(score_w = score_m) as draws,
group_concat(case when rn <= 3 then date end) as recent
from (
select t.*, row_number() over(order by date desc) rn
from `6dos7me3xn8` t
) t
The subquery ranks records by descending date; we can then use that information in the outer query. Note that you don't need the case expressions: MySQL evaluates true/false conditions as 1/0 in numeric contet.
In earlier versions, the simpler approach is probably a row-limiting subquery:
select
sum(score_w > score_m) as wins_w,
sum(score_m > score_w) as wins_m,
sum(score_w = score_m) as draws,
(select group_concat(date) from (select date from `6dos7me3xn8` order by date desc limit 3) t) as recent
from `6dos7me3xn8`
You are doing a cross join. TO get the most recent three dates, you can use:
FROM (SELECT mn.*, DENSE_RANK() OVER (ORDER BY date desc) as seqnum
FROM 6dos7me3xn8 mn
) mn
WHERE seqnum <= 3
You have no ON clause to specify the relationship between the subquery and the table you're joining it to. So you get a full cross product.
You also need ORDER BY to make it return the 3 most recent dates, not any 3 dates.
SELECT
SUM(CASE WHEN mn.score_w, 0) > mn.score_m THEN 1 ELSE 0 END) AS wins_w,
SUM(CASE WHEN mn.score_m > mn.score_w THEN 1 ELSE 0 END) AS wins_m,
SUM(CASE WHEN mn.score_w = mn.score_m THEN 1 ELSE 0 END) AS draws,
GROUP_CONCAT(jn.date) AS recent
FROM 6dos7me3xn8 mn
JOIN (
SELECT DISTINCT date
FROM 6dos7me3xn8
ORDER BY date DESC
LIMIT 3
) jn ON jn.date = mn.date
We have a table with data from different nodes and one of the column will have status report as "compliant or non-compliant", sample data as below
I want to filter the table in such a way that if any of the checks on a node shows non compliant, it should be flagged as non-compliant and rest as compliant. Using below query i am able to do it
SELECT COUNT(*) AS total_nodes,
SUM(fully_compliant = 0) AS Non_compliant_nodes,
SUM(fully_compliant = 1) AS compliant_nodes
FROM (
SELECT Node, CASE WHEN SUM(Status = 'Compliant') = COUNT(*) THEN 1 ELSE 0 END AS fully_compliant
FROM your_table GROUP BY Node
)
Now, i want to group and split the result by dept as below, how can i achieve this
I think you're looking for this:
select dept,
count(*) as total_nodes,
sum(case when non_compliant_chk = 0 then 1 else 0 end) as compliant_nodes,
sum(case when non_compliant_chk > 0 then 1 else 0 end) as non_compliant_nodes
from (
select dept,
node,
sum(case when 'Non-Compliant' then 1 else 0 end) as non_compliant_chk
from your_table
group by dept,
node
) v
group by dept;
With few modifications to what Brian suggested, I am able to get the desired result
select dept,
count(*) as total_nodes,
sum(case when non_compliant_chk = 0 then 1 else 0 end) as compliant_nodes,
sum(case when non_compliant_chk > 0 then 1 else 0 end) as non_compliant_nodes
from (
select dept,
node,
COUNT(CASE WHEN Compliance-Status = 'Non-Compliant' THEN 1 END) 'non_compliant_chk'
from table WHERE DOR >= DATE(NOW()) - INTERVAL 7 DAY
group by Dept,
Node
) v
group by Dept;
i have a query like below
select project_task_id,
status_id,
sum(case when StatusID=1 then 1 else 0 end) as task_id=1,
sum(case whenStatusID=2 then 1 else 0 end) as task_id=2,
sum(case when StatusID=3 then 1 else 0 end) as task_id=3,
sum(case when StatusID=4 then 1 else 0 end) as task_id=4,
sum(case when StatusID=5 then 1 else 0 end) as task_id=5,
sum(case when StatusID=6 then 1 else 0 end) as task_id=6,
sum(case when StatusID=7 then 1 else 0 end) as task_id=7,
from"Projects".work_unit_status
group by project_task_id,status_id;
I'm getting the below attached output:
https://i.stack.imgur.com/1wfD1.png
and I want to get the below expected output:
https://i.stack.imgur.com/Zql9z.png
include zero's if the status_id is blank
please any one help on this
Try this: use the addition of all sum column
select project_task_id,status_id,
isnull(sum(case when StatusID=1 then 1 else 0 end),0)+
isnull(sum(case whenStatusID=2 then 1 else 0 end),0) +
isnullsum(case when StatusID=3 then 1 else 0 end),0) +
isnullsum(case when StatusID=4 then 1 else 0 end),0)+
isnullsum(case when StatusID=5 then 1 else 0 end),0) +
isnullsum(case when StatusID=6 then 1 else 0 end),0) +
isnullsum(case when StatusID=7 then 1 else 0 end),0) as count_status
from"Projects".work_unit_status group by project_task_id,status_id
use in with your case
with t1 as (
select project_task_id,
status_id,
sum(case when StatusID in (1,2,3,4,5,6,7) then 1 else 0)
as sum_s
from "Projects".work_unit_status
group by project_task_id,status_id
) ,
t2 as
(
select * from (
select 1 as statusid
union
select 2
union
select 3
union
select 4
union
select 5
union
select 6
union
select 7 ) t
) select t1.project_task_id,
t2.statusid,
case when t1.sum_s>0 or not null
then sum_s else 0 end as total
t2 full join t1 on t2.statusid=t1.status_id
Without knowing the exact table structure I assumed that status_id and statusId refer to the same column. (If they are different columns, we need to use StatusId in the COUNT.)
Based on the expected output, you want to count the status_id and group by project_task_id. To make sure that every status is represented for every task, first, we need to create a subquery of all possible project_task_id/status_id combinations. Then we use that with the aggregate values of the original table.
select
ps.project_task_id,
ps.status_id,
count(w.status_id) as total
from (
select distinct
project_task_id,
s.status_id
from work_unit_status
cross join (select distinct status_id from work_unit_status) s
) ps
left join work_unit_status w
on ps.project_task_id = w.project_task_id and ps.status_id = w.status_id
group by
ps.project_task_id,
ps.status_id
If you really need to hardcode the statuses from 1 to 7, use the query below.
select
ps.project_task_id,
ps.status_id,
count(w.status_id) as total
from (
select distinct
project_task_id,
s.status_id
from work_unit_status
cross join (
select 1 as status_id
union select 2
union select 3
union select 4
union select 5
union select 6
union select 7
) s
) ps
left join work_unit_status w
on ps.project_task_id = w.project_task_id and ps.status_id = w.status_id
group by
ps.project_task_id,
ps.status_id
order by
ps.project_task_id,
ps.status_id
In SQL how do you show 0 if record is null?
select sales_id, totalbuy, totalsell, totalbuy + totalsell as total from
(select sales_id, SUM(CASE WHEN side= 'buy' THEN 1 ELSE 0 END) AS totalbuy,
SUM(CASE WHEN side= 'sell' THEN 1 ELSE 0 END) AS totalsell
from car_orders
where sales_id in ('sales1', 'sales2', 'sales3', 'sales4')only
GROUP BY sales_id)q
order by total desc
limit 0, 10;
After car_orders I have tried inserting*(car_orders+ISNULL(car_orders,0)) but get an error.
Building from Sohnee's answer, here is the SQL I think you want to use:
SELECT
sales_id,
IFNULL(totalbuy, 0),
IFNULL(totalsell, 0),
IFNULL(totalbuy, 0) + IFNULL(totalsell, 0) as total
FROM
(
SELECT
sid as sales_id,
SUM(CASE WHEN side = 'buy' THEN 1 ELSE 0 END) AS totalbuy,
SUM(CASE WHEN side = 'sell' THEN 1 ELSE 0 END) AS totalsell
FROM
( SELECT 'sales1' as sid UNION SELECT 'sales2' UNION SELECT 'sales3' UNION SELECT 'sales4' ) mysalesids
LEFT OUTER JOIN car_orders
ON sales_id = sid
GROUP BY
sales_id
) q
ORDER BY total DESC
LIMIT 0, 10;
The key to the above is the "LEFT OUTER JOIN". If you can have the 'sales1', 'sales2', 'sales3' values in their own table, that would be preferable rather than having a sub-select.
Hope this helps,
john...
When you use +, both arguments must be of the same type. I can't tell from your question what car is, but I assume it isn't compatible with a 0.
It is better to use CONCAT in these cases.
CONCAT(car, IFNULL(car_orders, 0))
If car_orders is a table, it isn't valid here - you must use a column, i.e. car_orders.MyColumn
How To Use IFNULL
I don't think you can end up with a null here, given your query, buy you would use IFNULL like this:
SELECT
sales_id,
IFNULL(totalbuy, 0),
IFNULL(totalsell, 0),
IFNULL(totalbuy, 0) + IFNULL(totalsell, 0) as total
FROM
(
SELECT
sales_id,
SUM(CASE WHEN side = 'buy' THEN 1 ELSE 0 END) AS totalbuy,
SUM(CASE WHEN side = 'sell' THEN 1 ELSE 0 END) AS totalsell
FROM
car_orders
WHERE
sales_id in ('sales1', 'sales2', 'sales3', 'sales4')
GROUP BY
sales_id
) q
ORDER BY total DESC
LIMIT 0, 10;