Currently I have an orders table that is formatted with a row per month:
id
order_month
order_count
order_sum
111
2021-07
5
50
111
2021-08
10
50
111
2021-09
1
100
222
2021-07
8
80
222
2021-08
2
50
222
2021-09
1
80
Is there a way to format the SQL query so that the ouput has 1 row per id, and the other values are added as columns? E.g. something like:
id
2021-07_order_count
2021-07_order_sum
2021-08_order_count
2021-08_order_sum
2021-09_order_count
2021-09_order_sum
111
5
50
10
50
1
100
222
8
80
2
50
1
80
I think I am close with the following query:
SELECT
merchant_id,
(CASE WHEN order_month = '2021-07' THEN order_count ELSE 0 END) as '2021-07-orderCount',
(CASE WHEN order_month = '2021-07' THEN order_sum ELSE 0 END) as '2021-07-orderSum',
(CASE WHEN order_month = '2021-08' THEN order_count ELSE 0 END) as '2021-08-orderCount',
(CASE WHEN order_month = '2021-08' THEN order_sum ELSE 0 END) as '2021-08-orderSum',
(CASE WHEN order_month = '2021-09' THEN order_count ELSE 0 END) as '2021-09-orderCount',
(CASE WHEN order_month = '2021-09' THEN order_sum ELSE 0 END) as '2021-09-orderSum'
FROM orders
ORDER BY id
It is creating a separate column and putting the correct values in each column.
However when I try and group by Id it then only shows the first result:
Thank you.
You need conditional aggregation:
SELECT id,
MAX(CASE WHEN order_month = '2021-07' THEN order_count ELSE 0 END) `2021-07-orderCount`,
MAX(CASE WHEN order_month = '2021-07' THEN order_sum ELSE 0 END) `2021-07-orderSum`,
MAX(CASE WHEN order_month = '2021-08' THEN order_count ELSE 0 END) `2021-08-orderCount`,
MAX(CASE WHEN order_month = '2021-08' THEN order_sum ELSE 0 END) `2021-08-orderSum`,
MAX(CASE WHEN order_month = '2021-09' THEN order_count ELSE 0 END) `2021-09-orderCount`,
MAX(CASE WHEN order_month = '2021-09' THEN order_sum ELSE 0 END) `2021-09-orderSum`
FROM orders
GROUP BY id
ORDER BY id;
See the demo.
Related
The table that needs to be queried looks like this:
ID
UserID
ProductCodes
1
33
9999
2
456
3051
3
456
9999
4
456
3051
4
33
9999
How would I write a SQL query to find out which users have at least one productCodes = '9999' and also have more than 10 productCodes <> '9999'?
You can use GROUP BY and HAVING:
SELECT
UserID
FROM dbo.YourTable
GROUP BY
UserId
HAVING SUM(CASE WHEN ProductCodes = '9999' THEN 1 ELSE 0 END) >= 1
AND COUNT(DISTINCT ProductCodes) >= 11
;
Use Case or Intersect (case is more performant)
SELECT UserID, SUM (case when ProductCodes='9999' then 1 else 0 end) PC9999
, SUM (case when ProductCodes<>'9999' then 1 else 0 end) PCNot9999
FROM dbo.Users
WHERE ProductCodes='9999'
GROUP BY UserID
HAVING SUM (case when ProductCodes='9999' then 1 else 0 end)>0
AND SUM (case when ProductCodes<>'9999' then 1 else 0 end) >10
I ended up going with this. It allows us to get specific with how many times a '9999' product code has been used in comparison with other codes.
SELECT
UserID
FROM Session_Hst
GROUP BY
UserID
HAVING SUM(CASE WHEN ProductCodes = '9999' THEN 1 ELSE 0 END) >= 1
AND COUNT(CASE WHEN ProductCodes <> '9999' THEN 1 ELSE null END ) >= 10
;
I have a column containing date ranges and the number of days passed associated to a specific ID (one to many), based on the number of records associated to it, I want those results split into columns instead of individual rows, so from this:
id_hr dd beg end
----------------------------------------
1 10 05/01/2019 15/01/2019
1 5 03/02/2019 08/02/2019
2 8 07/03/2019 15/03/2019
Could become this:
id_hr dd beg end dd beg end
--------------------------------- ---------------------
1 10 05/01/2019 15/01/2019 5 03/02/2019 08/02/2019
2 8 07/03/2019 15/03/2019
I did the same in a worksheet (pivot table) but the table became as slow as it could get, so I'm looking for a more friendly approach in SQL, I did a CTE which number the associated rows and then select each one and display them in new columns.
;WITH CTE AS(
SELECT PER_PRO, ID_HR, NOM_INC, rut_dv, dias_dur, INI, FIN,
ROW_NUMBER()OVER(PARTITION BY ID_HR ORDER BY SUBIDO) AS RN
FROM dbo.inf_vac WHERE PER_PRO = 201902
)
SELECT ID_HR, NOM_INC, rut_dv,
(case when rn = 1 then DIAS_DUR end) as DIAS_DUR1,
(case when rn = 1 then INI end) as INI1,
(case when rn = 1 then FIN end) as FIN1,
(case when rn = 2 then DIAS_DUR end) as DIAS_DUR2,
(case when rn = 2 then INI end) as INI2,
(case when rn = 2 then FIN end) as FIN2,
(case when rn = 3 then DIAS_DUR end) as DIAS_DUR3,
(case when rn = 3 then INI end) as INI3,
(case when rn = 3 then FIN end) as FIN3
FROM CTE
Which gets me each column on where it should be but not grouped. Using GROUP BY displays an error on the CTE select.
rn id_hr dd beg end dd beg end
----------------------------------- ------------------------
1 1 10 05/01/2019 15/01/2019 NULL NULL NULL
2 1 NULL NULL NULL 5 03/02/2019 08/02/2019
1 2 8 07/03/2019 15/03/2019 NULL NULL NULL
Is there any way to group them on the second select?
You have additional columns in the result set that are not in the query. However, this should work:
SELECT ID_HR,
max(case when rn = 1 then DIAS_DUR end) as DIAS_DUR1,
max(case when rn = 1 then INI end) as INI1,
max(case when rn = 1 then FIN end) as FIN1,
max(case when rn = 2 then DIAS_DUR end) as DIAS_DUR2,
max(case when rn = 2 then INI end) as INI2,
max(case when rn = 2 then FIN end) as FIN2,
max(case when rn = 3 then DIAS_DUR end) as DIAS_DUR3,
max(case when rn = 3 then INI end) as INI3,
max(case when rn = 3 then FIN end) as FIN3
FROM CTE
GROUP BY ID_HR;
Yes, you can GROUP BY all the non-CASE columns, and apply MAX to each of the CASE-expression columns.
I have a Tickets table in My database , each Ticket have a status_id (1,2,3)
1: Ticket IN PROGRESS
2: Ticket Out Of time
3: Ticket Closed
I want using SQL to calculate the number of tickets for each status .
Calculate the cumulative total for each Status in a specific Date, I have already a column affectation_Date that contains the date where the status of ticket has been changed .
Use conditional aggregation as
SELECT TicketID,
AffectationDate,
SUM(CASE WHEN StatusID = 1 THEN 1 ELSE 0 END) InProgress,
SUM(CASE WHEN StatusID = 2 THEN 1 ELSE 0 END) OuOfTime,
SUM(CASE WHEN StatusID = 3 THEN 1 ELSE 0 END) Closed,
COUNT(1) Total
FROM Tickets
GROUP BY TicketID,
AffectationDate
ORDER BY TicketID,
AffectationDate;
Or if you want to GROUP BY AffectationDate only
SELECT AffectationDate,
SUM(CASE WHEN StatusID = 1 THEN 1 ELSE 0 END) TotalInProgress,
SUM(CASE WHEN StatusID = 2 THEN 1 ELSE 0 END) TotalOutOfTime,
SUM(CASE WHEN StatusID = 3 THEN 1 ELSE 0 END) TotalClosed,
COUNT(1) TotalStatusThisDate
FROM Tickets
GROUP BY AffectationDate
ORDER BY AffectationDate;
Live Demo
Using conditional counts.
SELECT affectation_Date,
COUNT(CASE WHEN status_id = 1 THEN 1 END) AS TotalInProgress,
COUNT(CASE WHEN status_id = 2 THEN 1 END) AS TotalOutOfTime,
COUNT(CASE WHEN status_id = 3 THEN 1 END) AS TotalClosed
FROM Tickets t
GROUP BY affectation_Date
ORDER BY affectation_Date
you may use the desired filter condition for the date criteria
SELECT COUNT(1), STATUS
FROM tickets
WHERE affectation_Date >= 'someDate'
group by status
Regards
You just need to group by status and count the number of tickets in each group:
select status, count(*) as number
from Tickets
where dt >= '2019-01-01 00:00:00' and dt < '2019-01-02 00:00:00'
group by status
having status >= 1 and status <= 3
This adds the Cumulative Sum to the existing answers:
SELECT AffectationDate,
Sum(CASE WHEN StatusID = 1 THEN 1 ELSE 0 END) AS TotalInProgress,
Sum(CASE WHEN StatusID = 2 THEN 1 ELSE 0 END) AS TotalOutOfTime,
Sum(CASE WHEN StatusID = 3 THEN 1 ELSE 0 END) AS TotalClosed,
Count(*) as TotalStatusThisDate,
Sum(Sum(CASE WHEN StatusID = 1 THEN 1 ELSE 0 END)) Over (ORDER BY AffectationDate) AS cumTotalInProgress,
Sum(Sum(CASE WHEN StatusID = 2 THEN 1 ELSE 0 END)) Over (ORDER BY AffectationDate) AS cumTotalOutOfTime,
Sum(Sum(CASE WHEN StatusID = 3 THEN 1 ELSE 0 END)) Over (ORDER BY AffectationDate) AS cumTotalClosed,
Sum(Count(*)) Over (ORDER BY AffectationDate) AS cumTotalStatusThisDate
FROM Tickets
GROUP BY AffectationDate
ORDER BY AffectationDate;
I am trying to generate a report, and so far have one completed that gives me how many orders, for each day, are in status 1-9.
TableA structure looks like this:
Sales Order | Order Status | Order Date
123456789 | 1 | 2017-02-22 00:00:00.000
123456790 | 0 | 2017-02-21 00:00:00.000
TableB structure looks like this:
Sales Order | Price
123456789 | 123.00
123456789 | 42.00
123456790 | 56.00
123456790 | 28.00
This code:
SELECT
MAX(year([OrderDate])) as Yr,
MAX(MONTH([OrderDate])) as M,
Day([OrderDate]) as Day,
sum(case when [OrderStatus]='0' THEN 1 ELSE 0 END) AS 'STATUS"0"',
sum(case when [OrderStatus]='1' THEN 1 ELSE 0 END) AS 'STATUS"1"',
sum(case when [OrderStatus]='2' THEN 1 ELSE 0 END) AS 'STATUS"2"',
sum(case when [OrderStatus]='4' THEN 1 ELSE 0 END) AS 'STATUS"4"',
sum(case when [OrderStatus]='8' THEN 1 ELSE 0 END) AS 'STATUS"8"',
sum(case when [OrderStatus]='9' THEN 1 ELSE 0 END) AS 'STATUS"9"',
sum(case when [OrderStatus]='S' THEN 1 ELSE 0 END) AS 'STATUS"S"',
sum(case when [OrderStatus]='*' THEN 1 ELSE 0 END) AS 'STATUS"*"',
sum(case when [OrderStatus]='/' THEN 1 ELSE 0 END) AS 'STATUS"/"'
FROM
SorMaster
WHERE
YEAR([OrderDate]) = YEAR(GETDATE())
GROUP BY
DATENAME(month, DateAdd(month, Month([OrderDate]) - 1, Cast('2008-01-01' AS Datetime))), Day([OrderDate])
ORDER BY
Yr DESC, M DESC, Day DESC
Returns this:
Yr | M | Day | STATUS"0" | STATUS"1" | STATUS"2" | STATUS"4" | STATUS"8" | STATUS"9" | STATUS"S" | STATUS"*" | STATUS"/"
2017 2 22 0 2 0 1 0 0 5 0 0
2017 2 21 0 0 0 7 0 0 0 0 0
This is PERFECT for my first report.
Now, comes the trouble. My Problem is that I need to nest-query Table B, so that instead of returning a count(orders), I need the sum(orders) those totals for each order are in Table B.
Using the above example, the query would need to return something like this:
Yr | M | Day | STATUS"0" | STATUS"1" | STATUS"2" | STATUS"4" | STATUS"8" | STATUS"9" | STATUS"S" | STATUS"*" | STATUS"/"
2017 2 22 0 165 0 0 0 0 0 0 0
2017 2 21 84 0 0 0 0 0 0 0 0
Any pointers?
Just join to TableB:
SELECT MAX(year([t1.OrderDate])) AS Yr,
MAX(MONTH([t2.OrderDate])) AS M,
DAY([t1.OrderDate]) AS Day,
SUM(CASE WHEN [OrderStatus] = '0' THEN t2.Price ELSE 0 END) AS 'STATUS"0"',
SUM(CASE WHEN [OrderStatus] = '1' THEN t2.Price ELSE 0 END) AS 'STATUS"1"',
SUM(CASE WHEN [OrderStatus] = '2' THEN t2.Price ELSE 0 END) AS 'STATUS"2"',
SUM(CASE WHEN [OrderStatus] = '4' THEN t2.Price ELSE 0 END) AS 'STATUS"4"',
SUM(CASE WHEN [OrderStatus] = '8' THEN t2.Price ELSE 0 END) AS 'STATUS"8"',
SUM(CASE WHEN [OrderStatus] = '9' THEN t2.Price ELSE 0 END) AS 'STATUS"9"',
SUM(CASE WHEN [OrderStatus] = 'S' THEN t2.Price ELSE 0 END) AS 'STATUS"S"',
SUM(CASE WHEN [OrderStatus] = '*' THEN t2.Price ELSE 0 END) AS 'STATUS"*"',
SUM(CASE WHEN [OrderStatus] = '/' THEN t2.Price ELSE 0 END) AS 'STATUS"/"'
FROM SorMaster t1
LEFT JOIN TableB t2
ON t1.[Sales Order] = t2.[Sales Order]
WHERE YEAR([OrderDate]) = YEAR(GETDATE())
GROUP BY DATENAME(month,DateAdd(month,Month([OrderDate])-1,Cast('2008-01-01' AS Datetime))),
DAY([OrderDate])
ORDER BY Yr DESC, M DEACLLSC, Day DESC
That's not too difficult, just a matter of LEFT JOINing in table B and then summing the prices in that. There's a couple of small tricks here. You want to LEFT JOIN to ensure that rows in table A always show up, even if there are no corresponding rows in table B. Secondly, in your SUM() statement, you'll need to add a COALESCE(...,0.00) to ensure you're summing decimals and no NULL values creep in from the LEFT JOIN. Oddly in databases, NULL + {anything} = NULL.
For the below query, you'll need to change the name of TableB to whatever the table name is, and the JOIN predicate will need to have the column names named accurately, and delimited correctly if they contain spaces. For example, in MSSQL the delimiters are [ and ], e.g. MyTable.[My Column With Spaces]
SELECT
MAX(YEAR([OrderDate])) as Yr,
MAX(MONTH([OrderDate])) as M,
DAY([OrderDate]) as Day,
sum(case when [OrderStatus]='0' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"0"',
sum(case when [OrderStatus]='1' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"1"',
sum(case when [OrderStatus]='2' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"2"',
sum(case when [OrderStatus]='4' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"4"',
sum(case when [OrderStatus]='8' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"8"',
sum(case when [OrderStatus]='9' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"9"',
sum(case when [OrderStatus]='S' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"S"',
sum(case when [OrderStatus]='*' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"*"',
sum(case when [OrderStatus]='/' THEN COALESCE(TableB.Price, 0.00) ELSE 0.00 END) AS 'STATUS"/"'
FROM SorMaster
LEFT OUTER JOIN TableB
ON TableB.SalesOrder = SorMaster.SalesOrder
WHERE YEAR([OrderDate]) = YEAR(GETDATE())
GROUP BY
DATENAME(month,DateAdd(month,Month([OrderDate])-1,Cast('2008-01-01' AS Datetime))),
DAY([OrderDate])
ORDER BY
Yr DESC,
M DESC,
Day DESC
By (left) joining tableB and replacing your 'count' ones by tableB.price you should get the sum of all positions of orders with the according status.
I have a huge database with over 4 million rows that look like that:
Customer ID Shop
1 Asda
1 Sainsbury
1 Tesco
2 TEsco
2 Tesco
I need to count customers that within last 4 weeks had shopped in all 3 shops Tesco Sainsbury and Asda. Can you please advice if its possible to do it with subqueries?
This is an example of a "set-within-sets" subquery. You can solve it with aggregation:
select customer_id
from Yourtable t
where <shopping date within last four weeks>
group by customer_id
having sum(case when shop = 'Asda' then 1 else 0 end) > 0 and
sum(case when shop = 'Sainsbury' then 1 else 0 end) > 0 and
sum(case when shop = 'Tesco' then 1 else 0 end) > 0;
This structure is quite flexible. So if you wanted Asda and Tesco but not Sainsbury, then you would do:
select customer_id
from Yourtable t
where <shopping date within last four weeks>
group by customer_id
having sum(case when shop = 'Asda' then 1 else 0 end) > 0 and
sum(case when shop = 'Sainsbury' then 1 else 0 end) = 0 and
sum(case when shop = 'Tesco' then 1 else 0 end) > 0;
EDIT:
If you want a count, then use this as a subquery and count the results:
select count(*)
from (select customer_id
from Yourtable t
where <shopping date within last four weeks>
group by customer_id
having sum(case when shop = 'Asda' then 1 else 0 end) > 0 and
sum(case when shop = 'Sainsbury' then 1 else 0 end) > 0 and
sum(case when shop = 'Tesco' then 1 else 0 end) > 0
) t