SQL Query: Cannot perform aggregate functions on sub queries - sql

I have the following SQL query
SELECT
[Date],
DATENAME(dw,[Date]) AS Day,
SUM(CASE WHEN ChargeCode IN (SELECT ChargeCode FROM tblChargeCodes WHERE Chargeable = 1) THEN Units ELSE 0 END) ChargeableTotal,
SUM(CASE WHEN ChargeCode IN (SELECT ChargeCode FROM tblChargeCodes WHERE Chargeable = 0) THEN Units ELSE 0 END) NotChargeableTotal,
SUM(Units) AS TotalUnits
FROM
tblTimesheetEntries
WHERE
UserID = 'PJW'
AND Date >= '2013-01-01'
GROUP BY
[Date]
ORDER BY
[Date] DESC;
But I get the error message:
Cannot perform an aggregate function on an expression containing an aggregate or a subquery.
Because I am using sub queries in the Case Else Summation.
How can I revise my query to get 2 x Sums of [Units] one for Chargeable = true, and one for Chargeable = false, even though the Chargeable field is in a different table to all the other information. The two tables are linked by ChargeCode which appears in both tblTimesheetEntries and tblChargeCodes.

Have you tried joining the tables on the chargeCode:
SELECT e.[Date],
DATENAME(dw,e.[Date]) AS Day,
SUM(CASE WHEN c.Chargeable = 1 THEN e.Units ELSE 0 END) ChargeableTotal,
SUM(CASE WHEN c.Chargeable = 0 THEN e.Units ELSE 0 END) NotChargeableTotal,
SUM(e.Units) AS TotalUnits
FROM tblTimesheetEntries e
LEFT JOIN tblChargeCodes c
on e.ChargeCode = c.ChargeCode
WHERE e.UserID = 'PJW'
AND e.Date >= '2013-01-01'
GROUP BY e.[Date]
ORDER BY e.[Date] DESC;

Related

A SQL query for the retrieval of result based on input month

Table 1
Table 2
My requirement is to input the Redemption month and list the tickets that has been scanned double or more.
For example Ticket No. T1 has been scanned 2 times under pickup,only once under PickupOutforDelivery and 2 times under Delivery.
Result needed like this:
How can I write a query to get the result like this?
Tried:
SELECT
Ticket,
COUNT(Scantype = 0) AS Pickup,
COUNT(Scantype = 1) AS PickupOutforDelivery,
COUNT(Scantype = 2) AS Delivery
FROM
Scans
GROUP BY
Ticket, ScanType
HAVING
(Pickup > 1 OR PickupOutforDelivery > 1 OR Delivery > 1)
OR (Pickup >= 1 AND PickupOutforDelivery >= 1)
ORDER BY
Ticket
Result
Assuming that RedemptionMonth has a datatype of DATE (which is clearly required); the following query will give you the result you want, except the "cosmetic" part (breaking by year month for the report part) that you have to do on your application:
SELECT YEAR(RedemptionMonth) AS [YEAR], MONTH(RedemptionMonth) AS [MONTH], TicketNo,
COALESCE(SUM(CASE WHEN ScanName = 'Pickup' THEN 1 ELSE 0 END), 0) AS Pickup,
COALESCE(SUM(CASE WHEN ScanName = 'PickupOutForDelivery' THEN 1 ELSE 0 END), 0) AS PickupOutForDelivery ,
COALESCE(SUM(CASE WHEN ScanName = 'Delivery' THEN 1 ELSE 0 END), 0) AS Delivery
FROM [Table 1] AS T1
JOIN [Table 2] AS T2
ON T1.ScanType = T2.ScanType
GROUP BY YEAR(RedemptionMonth) AS [YEAR], MONTH(RedemptionMonth) AS [MONTH], TicketNo
Because you are using the numeric value of scanType, no JOIN is needed. So, the only fix is needed for conditional aggregation:
SELECT Ticket,
SUM(CASE WHEN Scantype = 0 THEN 1 ELSE 0 END) as Pickup,
SUM(CASE WHEN Scantype = 1 THEN 1 ELSE 0 END) as PickupOutforDelivery,
SUM(CASE WHEN Scantype = 2 THEN 1 ELSE 0 END) as Delivery
FROM Scans
WHERE redemptionMonth = 'Jan-21'
GROUP BY Ticket
HAVING Pickup > 1 OR
PickupOutforDelivery > 1 OR
Delivery > 1 OR
(Pickup >= 1 AND PickupOutforDelivery >= 1)
ORDER BY Ticket;
Note that you can add redemptionMonth to the GROUP BY (and SELECT) to get the results for each month.
If redemptionMonth is really a date and not a string, then define the time period using a range of dates:
WHERE redemptionMonth >= '2021-01-01' AND
redemptionMonth < '2021-02-01'

How to remove rows that are all zeros from SQL query

Should be simple to do this but cannot make it work. Have also looked at other similar answers. This code generates a report. There are 3 primary categories: General Adjustments, Finance Charges, and Bad Debts. What I am trying to accomplish...
If these 3 columns are all zeros in a row, delete that row. That's it.
DECLARE #startdate DATETIME
DECLARE #lastdate DATETIME
SET #startdate = '20170701'
SET #lastdate = '20170731'
CREATE TABLE #a
(
ClientName VARCHAR(50),
JobName VARCHAR(50),
JobPartner VARCHAR(40),
[Date] DATETIME,
GenAdj MONEY,
FC MONEY,
BadDebt MONEY,
nomid INT
)
SELECT
C.searchname AS 'CLIENT NAME',
n.date AS 'DATE',
SUM(CASE WHEN NOMTYPEID=3 AND NOMSUBTYPEID=2 THEN TOTAL else 0 END) AS 'GENERAL ADJUSTMENTS',
SUM(CASE WHEN NOMTYPEID=3 AND NOMSUBTYPEID=1 THEN TOTAL else 0 END) AS 'FINANCE CHARGES',
SUM(CASE WHEN NOMTYPEID=3 AND NOMSUBTYPEID=0 THEN TOTAL else 0 END) AS 'BAD DEBTS',
sl.serviceline AS 'DEPARTMENT',
p.name AS 'PARTNER NAME'
FROM
tblNominal N
JOIN
tblNominalAccs A ON N.ContNominalID = A.NominalID
JOIN
tblclient C on N.clientid = C.clientid
LEFT JOIN
tbljob j ON n.jobid = j.JobID
LEFT JOIN
tblpartner p ON j.PartnerID = p.PartnerID
LEFT JOIN
tbljobtype jt ON j.jobtypeid = jt.jobtypeid
LEFT JOIN
tblserviceline sl ON j.servicelineid = sl.servicelineid
WHERE
Date >= #startdate AND DATE <= #LASTDATE
GROUP BY
C.searchname, n.date, sl.serviceline, p.name
ORDER BY
'CLIENT NAME' asc, 'FINANCE CHARGES' asc
DROP TABLE #a
I think I may be overthinking this. How can I delete the rows with all zeros?
Thank you for reading.
You need to use a HAVING clause after your GROUP BY but before your ORDER BY:
HAVING SUM(CASE WHEN NOMTYPEID=3 AND NOMSUBTYPEID=2 THEN TOTAL else 0 END) > 0
OR SUM(CASE WHEN NOMTYPEID=3 AND NOMSUBTYPEID=1 THEN TOTAL else 0 END) > 0
OR SUM(CASE WHEN NOMTYPEID=3 AND NOMSUBTYPEID=0 THEN TOTAL else 0 END) > 0
HAVING is kind of like a WHERE for group results.

How to show 0 value using COUNT and SELECTon a SQL query

I have ONLY 1 table called Meeting that stores all meeting requests.
This table can be EMPTY.
It has several columns including requestType (which can only be "MT") meetingStatus (can only be either pending, approved, denied or canceled) and meetingCreatedTime
I want to count how many requests of each status's type (in other words how many requests are pending, how many are approved, denied and canceled) for the last 30 days
Problem is that if there is no request then nothing display but I want to display 0, how do I do it? Here is my query now:
SELECT [requestType],
( SELECT COUNT ([requestType]) FROM [Meeting] WHERE CAST([meetingCreatedTime] AS DATE) >= CAST(DateAdd(DAY,-30,Getdate()) AS DATE) AND [meetingStatus] = 'Approved') As 'Approved',
( SELECT COUNT ([requestType]) FROM [Meeting] WHERE CAST([meetingCreatedTime] AS DATE) >= CAST(DateAdd(DAY,-30,Getdate()) AS DATE) AND [meetingStatus] = 'Pending') As 'Pending',
( SELECT COUNT ([requestType]) FROM [Meeting] WHERE CAST([meetingCreatedTime] AS DATE) >= CAST(DateAdd(DAY,-30,Getdate()) AS DATE) AND [meetingStatus] = 'Canceled') As 'Canceled',
( SELECT COUNT ([requestType]) FROM [Meeting] WHERE CAST([meetingCreatedTime] AS DATE) >= CAST(DateAdd(DAY,-30,Getdate()) AS DATE) AND [meetingStatus] = 'Denied') As 'Denied'
FROM [Meeting]
WHERE CAST([meetingCreatedTime] AS DATE) >= CAST(DateAdd(DAY,-30,Getdate()) AS DATE) GROUP BY [requestType]
Result:
What I want is:
SELECT
RT.requestType,
SUM(CASE WHEN M.meetingStatus = 'Approved' THEN 1 ELSE 0 END) AS Approved,
SUM(CASE WHEN M.meetingStatus = 'Pending' THEN 1 ELSE 0 END) AS Pending,
SUM(CASE WHEN M.meetingStatus = 'Canceled' THEN 1 ELSE 0 END) AS Canceled,
SUM(CASE WHEN M.meetingStatus = 'Denied' THEN 1 ELSE 0 END) AS Denied,
FROM
(SELECT DISTINCT requestType FROM Meeting) RT
LEFT OUTER JOIN Meeting M ON
M.requestType = RT.requestType AND
M.meetingCreatedTime >= DATEADD(DAY, -30, GETDATE())
GROUP BY
RT.requestType
The SUMs are a much clearer (IMO) and much more efficient way of getting the counts that you need. Using the requestType table (assuming that you have one) lets you get results for every request type even if there are no meetings of that type in the date range. The LEFT OUTER JOIN to the meeting table allows the request type to still show up even if there are no meetings for that time period.
All of your CASTs between date values seem unnecessary.
Move those subqueries into simple sum/case statements:
select rt.request_type,
sum(case when [meetingStatus] = 'Approved' then 1 else 0 end),
sum(case when [meetingStatus] = 'Pending' then 1 else 0 end),
sum(case when [meetingStatus] = 'Canceled' then 1 else 0 end),
sum(case when [meetingStatus] = 'Denied' then 1 else 0 end)
from ( select 'MT' ) rt (request_type) --hopefully you have lookup table for this
left
join [Meeting] m on
rt.request_type = m.request_type and
CAST([meetingCreatedTime] AS DATE) >= CAST(DateAdd(DAY,-30,Getdate()) AS DATE)
group
by rt.request_type;
This is one possible approach to force one line to be visible in any case. Adapt this to your needs...
Copy it into an empty query window and execute... play around with the WHERE part...
DECLARE #Test TABLE (ID INT IDENTITY, GroupingKey VARCHAR(100));
INSERT INTO #Test VALUES ('a'),('a'),('b');
SELECT TOP 1 tbl.CountOfA
,tbl.CountOfB
,tbl.CountOfC
FROM
(
SELECT 1 AS Marker
,(SELECT COUNT(*) FROM #Test WHERE GroupingKey='a') AS CountOfA
,(SELECT COUNT(*) FROM #Test WHERE GroupingKey='b') AS CountOfB
,(SELECT COUNT(*) FROM #Test WHERE GroupingKey='c') AS CountOfC
WHERE (1=1) --play here with (1=0) and (1=1)
UNION ALL
SELECT 2,0,0,0
) AS tbl
ORDER BY Marker

Revenue for two months date wise

I am trying to get data for last 2 month ...but the query does not give perfect result....
SELECT DAY(table_A.PaymentDate) as date1 ,
(case when MONTH(table_A.PaymentDate) = MONTH(CURRENT_TIMESTAMP) - 1
then CAST(SUM(table_A.Total_Amount) AS INT)
else 0
end) AS last_month_CNT,
(case when MONTH(table_A.PaymentDate) = MONTH(CURRENT_TIMESTAMP)
then CAST(SUM(table_A.Total_Amount) As INT)
else 0
end) as This_month_CNT
FROM Tbl_Pan_Paymentdetails table_A
FULL OUTER JOIN Tbl_Pan_Paymentdetails table_B
ON table_A.PaymentDate=table_B.PaymentDate
WHERE YEAR(table_A.PaymentDate) = YEAR(CURRENT_TIMESTAMP)
AND
table_A.PaymentDate >= DATEADD(MONTH, -1, GETDATE())
GROUP BY
DAY(table_A.PaymentDate) ,MONTH(table_A.PaymentDate)
order by
DAY(table_A.PaymentDate);
Move the entire case expression inside the sum function and don't include the month in the group by. Also, the full outer join seems unnecessary so I removed it.
This should be what you are looking for:
SELECT
DAY(PaymentDate) as date1 ,
SUM(CASE WHEN MONTH(PaymentDate) = MONTH(CURRENT_TIMESTAMP)-1 THEN CAST(Total_Amount AS INT) ELSE 0 END) AS last_month_CNT,
SUM(CASE WHEN MONTH(PaymentDate) = MONTH(CURRENT_TIMESTAMP) THEN CAST(Total_Amount AS INT) ELSE 0 END) AS This_month_CNT
FROM Tbl_Pan_Paymentdetails
WHERE YEAR(PaymentDate) = YEAR(CURRENT_TIMESTAMP)
AND PaymentDate >= DATEADD(MONTH, -1, GETDATE())
GROUP BY DAY(PaymentDate)
ORDER BY DAY(PaymentDate);

SQL Query: Including Two Sums based on different criteria from the same table

I currently have the following sql query
SELECT [Date], DATENAME(dw,[Date]) AS Day, SUM(Units) AS TotalUnits
FROM tblTimesheetEntries
WHERE UserID = 'PJW'
AND Date >= '2013-01-01'
GROUP BY [Date]
ORDER BY [Date] DESC;
Which returns the Total [Units] for a given user and date.
However I would like to produce two separate Total Units based on a new criteria i.e. whether or not the [Units] where 'Chargeable or Not'. There is a field in tblTimeSheets called Chargeable which is a boolean (true for chargeable, false for not). How do I incorporate this to show two Sums one for each type of Units?
You can use a CASE with the SUM() to calculate the separate totals:
SELECT [Date],
DATENAME(dw,[Date]) AS Day,
sum(case when Chargeable = 1 then Units else 0 end) ChargeableTotal,
sum(case when Chargeable = 0 then Units else 0 end) NotChargeableTotal,
sum(Units) AS TotalUnits
FROM tblTimesheetEntries
WHERE UserID = 'PJW'
AND Date >= '2013-01-01'
GROUP BY [Date]
ORDER BY [Date] DESC;