Aggregate Nested Pgsql query - postgresql

I'm trying to query a table to get me some results, but the way I'm doing it gives me the error: ERROR: aggregate function calls cannot be nested.
The table looks like this:
ID | canal_c1 | tarifacao | date | ativo_id
The query I'm trying is this:
SELECT
SUM(case when tarifacao = 'ForaPonta' then canal_c1 else 0 end) as ConsForaPonta,
MAX(case when tarifacao = 'ForaPonta' then canal_c1 else 0 end) as DemForaPonta,
ativo_id as ativo_id,
data_leitura_inicio::date as date
FROM
medicao
WHERE
medicao.ativo_id in (45) AND
medicao.tipo_leitura = 'Consumo' AND
medicao.data_leitura_inicio >= '2017-01-01' AND
medicao.data_leitura_inicio < '2017-01-10'
GROUP BY
medicao.ativo_id,
medicao.data_leitura_inicio::date
That gives me result like these:
query result
And that's fine what I need now is the datetime from the DemForaPonta field, in order to do that I trying this, but got that error.
MAX(case when tarifacao = 'ForaPonta' and
canal_c1 = MAX(case when tarifacao = 'ForaPonta' then canal_c1 else 0 end)
then data_leitura_inicio end) as DateDemForaPonta
Do you know how I could achieve this?
Thanks.
Edit:
Here's an example, the query result is the intended result.
example

It would be helpful to have an example with real data. But anyway, you can do what you need with nested selects, among other options.
Nested SELECTs
SELECT *,
MAX(case when tarifacao = 'ForaPonta' and canal_c1 = DemForaPonta then data_leitura_inicio END)
AS DateDemForaPonta
FROM(
SELECT
SUM(case when tarifacao = 'ForaPonta' then canal_c1 else 0 end) as ConsForaPonta,
MAX(case when tarifacao = 'ForaPonta' then canal_c1 else 0 end) as DemForaPonta,
ativo_id as ativo_id,
data_leitura_inicio::date as date
FROM
medicao
WHERE
medicao.ativo_id in (45) AND
medicao.tipo_leitura = 'Consumo' AND
medicao.data_leitura_inicio >= '2017-01-01' AND
medicao.data_leitura_inicio < '2017-01-10'
GROUP BY
medicao.ativo_id,
medicao.data_leitura_inicio::date
)data
Since I don't have data to test it, don't really know if it does the trick, but it should.

Related

Selecting multiple counted fields in T-SQL query

I am stuck on a SQL query, where I need to count the number of different values in a field.
My query is:
SELECT
Area
, (SELECT count(TestResult) FROM TestRun WHERE TestResult = 'PASS' AND TestRun.CreatedDate > '2019-11-18 01:00:00') as [Passed]
, (SELECT count(TestResult) FROM TestRun WHERE TestResult = 'FAIL' AND TestRun.CreatedDate > '2019-11-18 01:00:00') as [Failed]
, (SELECT count(TestResult) FROM TestRun WHERE TestResult = 'NOTRUN' AND TestRun.CreatedDate > '2019-11-18 01:00:00') as [NotRun]
FROM TestRun
WHERE dbo.TestRun.CreatedDate > '2019-11-18 01:00:00'
GROUP BY dbo.TestRun.Area, TestRun.TestResult, TestRun.CreatedDate
But The Results I get back look like this, instead of individual values:
You can create your own fields using case and then sum them.
SELECT Area
, SUM(CASE WHEN TestResult = 'PASS' THEN 1 ELSE 0 END) AS Passed
, SUM(CASE WHEN TestResult = 'FAIL' THEN 1 ELSE 0 END) AS Failed
, SUM(CASE WHEN TestResult = 'NOTRUN' THEN 1 ELSE 0 END) AS NotRun
FROM TestRun
WHERE CreatedDate > '2019-11-18 01:00:00'
GROUP BY Area

Combine Two Queries in one SUM Case

I have two queries with exact same grouping but I dont seem to be able to combine them in a correct way.
Query1:
SELECT
WorkPeriods.Id AS Z_Number,
CONVERT(VARCHAR, (CONVERT(DATE, WorkPeriods.StartDate, 103)), 103) AS Z_Date,
SUM(CASE WHEN Payments.Name = 'Cash' THEN Payments.Amount ELSE 0 END) AS Cash_Payments,
COUNT(CASE WHEN Payments.Name = 'Cash' THEN 1 END) AS No_of_Tickets_Cash,
SUM(CASE WHEN Payments.Name = 'Credit Card' THEN Payments.Amount ELSE 0 END) AS Credit_Card_Payments,
COUNT(CASE WHEN Payments.Name = 'Credit Card' THEN 1 END) AS No_of_Tickets_Credit_Card
FROM
Payments, WorkPeriods
WHERE
Payments.Date BETWEEN WorkPeriods.StartDate AND WorkPeriods.EndDate
GROUP BY
WorkPeriods.Id, WorkPeriods.StartDate
Query 2:
SELECT
WorkPeriods.Id AS Z_Number,
CONVERT(VARCHAR, (CONVERT(DATE, WorkPeriods.StartDate, 103)), 103) AS Z_Date,
SUM(CASE WHEN Orders.CalculatePrice = 0 THEN Orders.Quantity * Orders.Price ELSE 0 END) AS Gifts_Amount,
SUM(CASE WHEN Orders.CalculatePrice = 0 THEN Orders.Quantity ELSE 0 END) AS No_of_Gift_Orders
FROM
Orders, WorkPeriods
WHERE
Orders.CreatedDateTime BETWEEN WorkPeriods.StartDate AND WorkPeriods.EndDate
GROUP BY
WorkPeriods.Id, WorkPeriods.StartDate
Any advice on how to continue? I have already tried merging them using all 3 tables and all sum-count conditions but the result I get is wrong. I need all results to appear on the same row. Attached are query results
You can't just join them all in the one query, as you will get incorrect values as soon as you get multiple orders or payments in the same workperiod.
You could use the current queries as sub queries, and full join them to get the result. By using full join you get any results that are only on one table and not the other.
Select ISNULL(Pay.Z_Number, Ord.Z_Number) As Z_Number,
ISNULL(Pay.Z_Date, Ord.Z_Date) as Z_Date,
Pay.CashPayments,
Pay.No_of_Tickets_Cash,
Ord.Gifts_Amount
--other fields as appropriate
FROM (
--Query 1 here
) AS Pay
FULL OUTER JOIN (
--Query 2 here
) as Ord ON Pay.Z_Number = Ord.Z_Number and Pay.Z_Date = Ord.Z_Date
Another way to do this, is to create one sub query that has the data from both payments and orders in it unioned together, and then sum the resulting list in the outer query.
Below sample query may be helpful
SELECT
MAIN_T.Z_Number
,MAIN_T.Z_Date
,T1.Cash_Payments
,T1.Credit_Card_Payments
,T1.No_of_Tickets_Cash
,T1.No_of_Tickets_Credit_Card
,T2.Gifts_Amount
,T2.No_of_Gift_Orders
FROM
(SELECT DISTINCT
WorkPeriods.Id AS Z_Number,
CONVERT(VARCHAR, (CONVERT(DATE, WorkPeriods.StartDate, 103)), 103) AS Z_Date
FROM
Payments, WorkPeriods
WHERE
Payments.Date BETWEEN WorkPeriods.StartDate AND WorkPeriods.EndDate ) MAIN_T
LEFT JOIN
(SELECT
WorkPeriods.Id AS Z_Number,
CONVERT(VARCHAR, (CONVERT(DATE, WorkPeriods.StartDate, 103)), 103) AS Z_Date,
SUM(CASE WHEN Payments.Name = 'Cash' THEN Payments.Amount ELSE 0 END) AS Cash_Payments,
COUNT(CASE WHEN Payments.Name = 'Cash' THEN 1 END) AS No_of_Tickets_Cash,
SUM(CASE WHEN Payments.Name = 'Credit Card' THEN Payments.Amount ELSE 0 END) AS Credit_Card_Payments,
COUNT(CASE WHEN Payments.Name = 'Credit Card' THEN 1 END) AS No_of_Tickets_Credit_Card
FROM
Payments, WorkPeriods
WHERE
Payments.Date BETWEEN WorkPeriods.StartDate AND WorkPeriods.EndDate
GROUP BY
WorkPeriods.Id, WorkPeriods.StartDate) T1
ON MAIN_T.Z_Number=T1.Z_Number AND MAIN_T.Z_Date=T1.Z_Date
LEFT JOIN
(SELECT
WorkPeriods.Id AS Z_Number,
CONVERT(VARCHAR, (CONVERT(DATE, WorkPeriods.StartDate, 103)), 103) AS Z_Date,
SUM(CASE WHEN Orders.CalculatePrice = 0 THEN Orders.Quantity * Orders.Price ELSE 0 END) AS Gifts_Amount,
SUM(CASE WHEN Orders.CalculatePrice = 0 THEN Orders.Quantity ELSE 0 END) AS No_of_Gift_Orders
FROM
Orders, WorkPeriods
WHERE
Orders.CreatedDateTime BETWEEN WorkPeriods.StartDate AND WorkPeriods.EndDate
GROUP BY
WorkPeriods.Id, WorkPeriods.StartDate) T2
ON MAIN_T.Z_Number=T2.Z_Number AND MAIN_T.Z_Date=T2.Z_Date

how to select top 10 without duplicates

Using SQL Server 2012
I need to select TOP 10 Producer based on a ProducerCode. But the data is messed up, users were entering same Producers just spelled differently and with the same ProducerCode.
So I just need TOP 10, so if the ProducerCode is repeating, I just want to pick the first one in a list.
How can I achieve that?
Sample of my data
;WITH cte_TopWP --T
AS
(
SELECT distinct ProducerCode, Producer,SUM(premium) as NetWrittenPremium,
SUM(CASE WHEN PolicyType = 'New Business' THEN Premium ELSE 0 END) as NewBusiness1,
SUM(CASE WHEN PolicyType = 'Renewal' THEN Premium ELSE 0 END) as Renewal1,
SUM(CASE WHEN PolicyType = 'Rewrite' THEN Premium ELSE 0 END) as Rewrite1
FROM ProductionReportMetrics
WHERE YEAR(EffectiveDate) = 2016 AND TransactionType = 'Policy' AND CompanyLine = 'Arch Insurance Company'--AND ProducerType = 'Wholesaler'
GROUP BY ProducerCode,Producer
)
,
cte_Counts --C
AS
(
SELECT distinct ProducerCode, ProducerName, COUNT (distinct ControlNo) as Submissions2,
SUM(CASE WHEN QuotedPremium IS NOT NULL THEN 1 ELSE 0 END) as Quoted2,
SUM(CASE WHEN Type = 'New Business' AND Status IN ('Bound','Cancelled','Notice of Cancellation') THEN 1 ELSE 0 END ) as NewBusiness2,
SUM(CASE WHEN Type = 'Renewal' AND Status IN ('Bound','Cancelled','Notice of Cancellation') THEN 1 ELSE 0 END ) as Renewal2,
SUM(CASE WHEN Type = 'Rewrite' AND Status IN ('Bound','Cancelled','Notice of Cancellation') THEN 1 ELSE 0 END ) as Rewrite2,
SUM(CASE WHEN Status = 'Declined' THEN 1 ELSE 0 END ) as Declined2
FROM ClearanceReportMetrics
WHERE YEAR(EffectiveDate)=2016 AND CompanyLine = 'Arch Insurance Company'
GROUP BY ProducerCode,ProducerName
)
SELECT top 10 RANK() OVER (ORDER BY NetWrittenPremium desc) as Rank,
t.ProducerCode,
c.ProducerName as 'Producer',
NetWrittenPremium,
t.NewBusiness1,
t.Renewal1,
t.Rewrite1,
c.[NewBusiness2]+c.[Renewal2]+c.[Rewrite2] as PolicyCount,
c.Submissions2,
c.Quoted2,
c.[NewBusiness2],
c.Renewal2,
c.Rewrite2,
c.Declined2
FROM cte_TopWP t --LEFT OUTER JOIN tblProducers p on t.ProducerCode=p.ProducerCode
LEFT OUTER JOIN cte_Counts c ON t.ProducerCode=c.ProducerCode
You should use ROW_NUMBER to fix your issue.
https://msdn.microsoft.com/en-us/library/ms186734.aspx
A good example of this is the following answer:
https://dba.stackexchange.com/a/22198
Here's the code example from the answer.
SELECT * FROM
(
SELECT acss_lookup.ID AS acss_lookupID,
ROW_NUMBER() OVER
(PARTITION BY your_distinct_column ORDER BY any_column_you_think_is_appropriate)
as num,
acss_lookup.product_lookupID AS acssproduct_lookupID,
acss_lookup.region_lookupID AS acssregion_lookupID,
acss_lookup.document_lookupID AS acssdocument_lookupID,
product.ID AS product_ID,
product.parent_productID AS productparent_product_ID,
product.label AS product_label,
product.displayheading AS product_displayheading,
product.displayorder AS product_displayorder,
product.display AS product_display,
product.ignorenewupdate AS product_ignorenewupdate,
product.directlink AS product_directlink,
product.directlinkURL AS product_directlinkURL,
product.shortdescription AS product_shortdescription,
product.logo AS product_logo,
product.thumbnail AS product_thumbnail,
product.content AS product_content,
product.pdf AS product_pdf,
product.language_lookupID AS product_language_lookupID,
document.ID AS document_ID,
document.shortdescription AS document_shortdescription,
document.language_lookupID AS document_language_lookupID,
document.document_note AS document_document_note,
document.displayheading AS document_displayheading
FROM acss_lookup
INNER JOIN product ON (acss_lookup.product_lookupID = product.ID)
INNER JOIN document ON (acss_lookup.document_lookupID = document.ID)
)a
WHERE a.num = 1
ORDER BY product_displayheading ASC;
You could do this:
SELECT ProducerCode, MIN(Producer) AS Producer, ...
GROUP BY ProducerCode

Count based on Or is not differentiating the count

My results are showing both counts the same but there should be some that have different counts as CarCode is sometimes null.
SELECT distinct car.carKey,
car.Weight,
car.CarCode,
COUNT(car.carKey)OVER(PARTITION BY car.carKey) AS TotalCarKeyCount,
COUNT(Case When (car.[Weight] IS not null) and (car.CarCode is null) as CarCountWithoutCode
then 0
else car.carKey End) OVER(PARTITION BY car.carKey) AS CarCount
from car
results show TotalCarKeyCount and CarCountWithoutCode always with the same counts like the case statement isn't working or something.
It sounds like you might want to use SUM() instead:
SELECT distinct car.carKey,
car.Weight,
car.CarCode,
COUNT(car.carKey)OVER(PARTITION BY car.carKey) AS TotalCarKeyCount,
SUM(Case When (car.[Weight] IS not null) and (car.CarCode is null) as CarCountWithoutCode
then 0 else 1 End) OVER(PARTITION BY car.carKey) AS CarCount
from car
SQL Fiddle demo showing the difference between using COUNT() and SUM():
create table test
(
id int
);
insert into test values
(1), (null), (23), (4), (2);
select
count(case when id is null then 0 else id end) [count],
sum(case when id is null then 0 else 1 end) [sum]
from test;
Count returns 5 and Sum returns 4. Or you can change the COUNT() to use null and the null values will be excluded in the final count()
select
count(case when id is null then null else id end) [count],
sum(case when id is null then 0 else 1 end) [sum]
from test;
Your query would be:
SELECT distinct car.carKey,
car.Weight,
car.CarCode,
COUNT(car.carKey)OVER(PARTITION BY car.carKey) AS TotalCarKeyCount,
COUNT(Case When (car.[Weight] IS not null) and (car.CarCode is null) as CarCountWithoutCode
then null else 1 End) OVER(PARTITION BY car.carKey) AS CarCount
from car
Change the then 0 to then null. Zero values are counted, nulls are not.

Help with difficult 'group by' clause

need some your help with a query.
I have a table Managers (ManagerId, ManagerName)
I have a table Statuses (StatusId, StatusName)
(There's about 10 statuses in that table)
I have a table Clients (ClientId, ClientName, ManagerId, StatusId, WhenAdded)
(WhenAdded is a datetime type)
It's obvious that field 'ManagerId' refers to a table 'Managers' and field 'StatusId' refers to a table 'Statuses'.
User wants to get some statistics about Managers over a period of time (from startDate to endDate using field 'WhenAdded') in the following table.
Columns:
ManagerName, NumberOfClients, NumberOfClientsWithStatus1, NumberOfClientsWithStatus2, NumberOfClientsWithStatus3 and so on.
Number of columns with name NumberOfClientsWithStatusI where i is a number of statuses equal to number of rows in table 'Statuses'.
How can I do that?
t-sql, sql server 2008 r2 express edition.
SELECT
ManagerName,
COUNT(*) AS NumberOfClients,
COUNT(CASE WHEN S.StatusId = 1 THEN 1 ELSE NULL END) AS NumberOfClientsWithStatus1,
COUNT(CASE WHEN S.StatusId = 2 THEN 1 ELSE NULL END) AS NumberOfClientsWithStatus2,
COUNT(CASE WHEN S.StatusId = 3 THEN 1 ELSE NULL END) AS NumberOfClientsWithStatus3,
...
FROM
Clients C
JOIN
Managers M ON C.ManagerId = M.ManagerId
JOIN
Statuses S ON C.StatusId = S.StatusId
WHERE
M.WhenAdded BETWEEN #startDate AND #endDate
GROUP BY
M.ManagerName
Note: there is no clean way to add arbritrary numbers of status columns in SQL (not just SQL Server) because its a fixed column output. You'd have to change the query for status, unless you deal with this in the client
Edit, after comment
SELECT
ManagerName,
COUNT(*) AS NumberOfClients,
COUNT(CASE WHEN S.StatusId = 1 THEN 1 ELSE NULL END) AS NumberOfClientsWithStatus1,
COUNT(CASE WHEN S.StatusId = 2 THEN 1 ELSE NULL END) AS NumberOfClientsWithStatus2,
COUNT(CASE WHEN S.StatusId = 3 THEN 1 ELSE NULL END) AS NumberOfClientsWithStatus3,
...
FROM
Managers M ON C.ManagerId = M.ManagerId
LEFT JOIN
Clients C
LEFT JOIN
Statuses S ON C.StatusId = S.StatusId
WHERE
M.WhenAdded BETWEEN #startDate AND #endDate
GROUP BY
M.ManagerName
If you know that statuses table will always contain a limited number of statuses, you can do this:
SELECT M.ManagerName,
COUNT(C.ClientId) NumberOfClients,
SUM(CASE WHEN S.StatusId= 1 THEN 1 ELSE 0 END) NumberOfClientsWithStatus1,
SUM(CASE WHEN S.StatusId= 2 THEN 1 ELSE 0 END) NumberOfClientsWithStatus2,
...
FROM Clients C
JOIN Managers M on M.ManagerId = C.ManagerId
JOIN Statuses S on S.StatusId = C.StatusId
WHERE C.WhenAdded BETWEEN startDate AND endDate
GROUP BY ManagerName