I have written a query in two models, but I can't figure out why the second query returns a greater number than the first one; while the number that the first one, COUNT(DISTINCT...) returns is correct:
WITH types(id) AS (VALUES('{1, 4, 5, 3}'::INTEGER[])),
date_gen64 AS
(
SELECT CAST (generate_series(date '10/1/2017', date '11/15/2017', interval
'1 day') AS date) as days ORDER BY days)
SELECT cl.class_date AS c_date,
count(DISTINCT (CASE WHEN co.id = 1 THEN p.id END)),
count(DISTINCT (CASE WHEN co.id = 2 THEN p.id END))
FROM person p
JOIN envelope e ON e.personID = p.id
JOIN "class" cl on cl.id = p.classID
JOIN course co ON co.id = cl.course_id AND co.id = 1
JOIN types ON cr.type_id = ANY (types.id)
RIGHT JOIN date_gen64 dg ON dg.days = cl.class_date
GROUP BY cl.class_date
ORDER BY cl.class_date
The above query returns 26 but following query returns 27!
The reason why I rewrote it with SUM is that the first query
was too slow. But my question is that why the second one counts more?
WITH types(id) AS (VALUES('{1, 4, 5, 3}'::INTEGER[]))
SELECT tmpcl.days,
SUM(CASE WHEN tmp80.course_id = 1 THEN 1
ELSE 0 END),
SUM(CASE WHEN tmp80.course_id = 2 THEN 1
ELSE 0 END)
FROM (
SELECT CAST (generate_series(date '10/1/2017', date '11/15/2017',
interval '1 day') AS date) as days ORDER BY days) tmpcl
LEFT JOIN (
SELECT DISTINCT p.id AS "person_id",
cl.class_date AS c_date,
co.id AS "course_id"
FROM person p
JOIN envelope e ON e.personID = p.id
JOIN "class" cl on cl.id = p.classID
JOIN course co ON co.id = cl.course_id
JOIN types ON cr.type_id = ANY (types.id)
WHERE co.id IN ( 1 , 2 )
) tmp80 ON tmpcl.days = tmp80.class_date
GROUP BY tmpcl.days
ORDER BY tmpcl.days
You can theoretically have multiple people enrolled in the same class on the same day. Indeed that would seem to be the main point of having classes. So each time there are multiple people assigned to the same class on the same day you can have a higher count than you would in your first query. Does that make sense?
You don't appear to be using p.id in that inner query so simply remove it and your counts should match.
WITH types(id) AS (VALUES('{1, 4, 5, 3}'::INTEGER[]))
SELECT tmpcl.days,
SUM(CASE WHEN tmp80.course_id = 1 THEN 1
ELSE 0 END),
SUM(CASE WHEN tmp80.course_id = 2 THEN 1
ELSE 0 END)
FROM (
SELECT CAST (generate_series(date '10/1/2017', date '11/15/2017',
interval '1 day') AS date) as days ORDER BY days) tmpcl
LEFT JOIN (
SELECT DISTINCT cl.class_date AS c_date,
co.id AS "course_id"
FROM person p
JOIN envelope e ON e.personID = p.id
JOIN "class" cl on cl.id = p.classID
JOIN course co ON co.id = cl.course_id
JOIN types ON cr.type_id = ANY (types.id)
WHERE co.id IN ( 1 , 2 )
) tmp80 ON tmpcl.days = tmp80.class_date
GROUP BY tmpcl.days
ORDER BY tmpcl.days
I have data something like this:
ID 1 1 1 1 1 1 1 1 1 1 1 1
Month J F M A M J J A S O N D
Status 1 0 0 1 0 1 0 0 1 1 1 1
ID 2 2 2 2 2 2 2 2 2 2 2 2
Month J F M A M J J A S O N D
Status 1 0 1 0 1 0 1 0 1 0 1 1
ID 3 3 3 3 3 3 3 3 3 3 3 3
Month J F M A M J J A S O N D
Status 0 0 0 0 0 0 0 0 0 0 0 1
Using t-SQL, I am trying to capture the month corresponding to the first STATUS = 1 in the last group of 1s for each ID, i.e., September, November and December in this example.
Here is the code I'm using:
IF OBJECT_ID('tempdb..#Temp1') IS NOT NULL DROP TABLE #Temp1
;WITH PARTITIONED1 AS
(SELECT t0.ID
, t0.Year_Month
, LAST_VALUE(t0.Year_Month) OVER (PARTITION BY t0.Account_Number ORDER BY t0.Year_Month) AS STATUS
, ROW_NUMBER() OVER (PARTITION BY t0.Account_Number ORDER BY t0.Year_Month) AS rn1
FROM #Temp0 t0
)
SELECT *
INTO #Temp1
FROM PARTITIONED1 p1
ORDER BY t0.ID
, t0.Year_Month
IF OBJECT_ID('tempdb..#Temp') IS NOT NULL DROP TABLE #Temp
SELECT *
INTO #Temp
FROM #Temp1 t1
WHERE t1.rn1 = (SELECT MAX(b.rn1) + 1 FROM #Temp1 b WHERE b.STATUS = 0)
GROUP BY t1.ID
, t1.Year_Month
, t1.rn1
However, this just returns the last instance where STATUS = 1 is achieved overall as the first 1 of the last group of 1s, in this case January.
I've tried using CASE statements and grouping in various combinations (hence the intermediate step reading the data into #Temp1), but have not been able to get results for all three IDs; is anyone able to assist?
Thanks in advance!
Assuming Ju for June and Jl for July:
--Sample Data
IF OBJECT_ID('tempdb..#Temp0') IS NOT NULL DROP TABLE #Temp0
CREATE TABLE #Temp0 (ID INT, Year_Month VARCHAR(1), Status INT)
INSERT INTO #Temp0
VALUES(1,'J',1),(1,'F',0),(1,'M',0),(1,'A',1),(1,'M',0),(1,'J',1),(1,'J',0),(1,'A',0),(1,'S',1),(1,'O',1),(1,'N',1),(1,'D',1),(2,'J',1),(2,'F',0),(2,'M',1),(2,'A',0),(2,'M',1),(2,'J',0),(2,'J',1),(2,'A',0),(2,'S',1),(2,'O',0),(2,'N',1),(2,'D',1),(3,'J',0),(3,'F',0),(3,'M',0),(3,'A',0),(3,'M',0),(3,'J',0),(3,'J',0),(3,'A',0),(3,'S',0),(3,'O',0),(3,'N',0),(3,'D',1);
--Query
WITH A
AS ( SELECT *,
CASE Year_Month
WHEN 'J' THEN 1
WHEN 'F' THEN 2
WHEN 'M' THEN 3
WHEN 'A' THEN 4
WHEN 'M' THEN 5
WHEN 'Ju' THEN 6
WHEN 'Jl' THEN 7
WHEN 'A' THEN 8
WHEN 'S' THEN 9
WHEN 'O' THEN 10
WHEN 'N' THEN 11
WHEN 'D' THEN 12
END
AS MonthNumber
FROM #Temp0 ),
StartingPoints
AS ( SELECT ID,
Year_Month,
MonthNumber,
Status
FROM A
WHERE NOT EXISTS
(
SELECT 1
FROM A
AS B
WHERE B.ID=A.ID
AND B.Status=A.Status-1
) ),
MonthRanking
AS ( SELECT A.*,
ROW_NUMBER( ) OVER( PARTITION BY A.ID ORDER BY A.MonthNumber )
AS rownum
FROM A
INNER JOIN
(
SELECT ID,
MAX( MonthNumber )+1
AS StartOfLastGroup
FROM StartingPoints
GROUP BY ID
)
AS B
ON A.ID=B.ID
AND A.MonthNumber>=B.StartOfLastGroup )
SELECT *
FROM MonthRanking
WHERE rownum=1;
Results:
If Month Names are recorded in Full as in July, June then this would work as well:
WITH StartingPoints
AS (SELECT ID,
Year_Month,
MonthNUmber = MONTH('01-'+Year_Month+'-2010'),
Status
FROM #Temp0
WHERE NOT EXISTS
(
SELECT 1
FROM #Temp0 AS B
WHERE B.ID = #Temp0.ID
AND B.Status = #Temp0.Status - 1
)),
MonthRanking
AS (SELECT A.*,
ROW_NUMBER() OVER(PARTITION BY A.ID ORDER BY MONTH('01-'+A.Year_Month+'-2010')) AS rownum
FROM #Temp0 AS A
INNER JOIN
(
SELECT ID,
MAX(MonthNumber) + 1 AS StartOfLastGroup
FROM StartingPoints
GROUP BY ID
) AS B ON A.ID = B.ID
AND MONTH('01-'+A.Year_Month+'-2010') >= B.StartOfLastGroup)
SELECT *
FROM MonthRanking
WHERE rownum = 1;
Results:
And if we assume that the data is as Iamdave assumes then it simply like so:
WITH StartingPoints
AS (SELECT ID,
Year_Month,
Status
FROM #Temp0
WHERE NOT EXISTS
(
SELECT 1
FROM #Temp0 AS B
WHERE B.ID = #Temp0.ID
AND B.Status = #Temp0.Status - 1
)),
MonthRanking
AS (SELECT A.*,
ROW_NUMBER() OVER(PARTITION BY A.ID ORDER BY Year_Month) AS rownum
FROM #Temp0 AS A
INNER JOIN
(
SELECT ID,
MAX(Year_Month) + 1 AS StartOfLastGroup
FROM StartingPoints
GROUP BY ID
) AS B ON A.ID = B.ID
AND A.Year_Month >= B.StartOfLastGroup)
SELECT *
FROM MonthRanking
WHERE rownum = 1;
Results:
You can do this with a couple derived tables that stack two window functions on top of one another (which can't be done in the same select). I have assumed that your data is slightly different to the table you have provided, based on the column names in your query. If they are not as I have them below, I strongly recommend having a look at how you store your data:
declare #t table(ID int, YearMonth int,StatusValue bit);
insert into #t values (1,201501,1),(1,201502,0),(1,201503,0),(1,201504,1),(1,201505,0),(1,201506,1),(1,201507,0),(1,201508,0),(1,201509,1),(1,201510,1),(1,201511,1),(1,201512,1),(2,201601,1),(2,201602,0),(2,201603,1),(2,201604,0),(2,201605,1),(2,201606,0),(2,201607,1),(2,201608,0),(2,201609,1),(2,201610,0),(2,201611,1),(2,201612,1),(3,201701,0),(3,201702,0),(3,201703,0),(3,201704,0),(3,201705,0),(3,201706,0),(3,201707,0),(3,201708,0),(3,201709,0),(3,201710,0),(3,201711,0),(3,201712,1);
with c as
(
select ID
,YearMonth
,StatusValue
,case when StatusValue = 1
and lead(StatusValue,1,1) over (partition by ID
order by YearMonth desc) = 0
then 1
else 0
end as c
from #t
), sc as
(
select ID
,YearMonth
,StatusValue
,sum(c) over (partition by ID order by YearMonth desc) as sc
from c
where c = 1
)
select ID
,YearMonth
,StatusValue
from sc
where sc = 1
order by ID;
Output:
+----+-----------+-------------+
| ID | YearMonth | StatusValue |
+----+-----------+-------------+
| 1 | 201509 | 1 |
| 2 | 201611 | 1 |
| 3 | 201712 | 1 |
+----+-----------+-------------+
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
Can someone help me with this query?
SELECT p.OwnerName, SUM(ru.MonthlyRent) AS PotentinalRent, SUM(
(SELECT COUNT(t.ID) * ru.MonthlyRent FROM tblTenant t
WHERE t.UnitID = ru.ID)
) AS ExpectedRent
FROM tblRentalUnit ru
LEFT JOIN tblProperty p ON p.ID = ru.PropertyID
GROUP BY p.OwnerName
I'm having problems with the second sum, it won't let me do it. Evidently SUM won't work on subqueries, but I need to calculate the expected rent (MonthlyRent if there is a tenant assigned to the RentalUnit's id, 0 of they're not). How can I make this work?
SELECT p.OwnerName, SUM(ru.MonthlyRent) AS PotentialRent, SUM(cnt) AS ExpectedRent
FROM tblRentalUnit ru
LEFT JOIN
tblProperty p
ON p.ID = ru.PropertyID
OUTER APPLY
(
SELECT COUNT(t.id) * ru.MonthlyRent AS cnt
FROM tblTenant t
WHERE t.UnitID = ru.ID
) td
GROUP BY p.OwnerName
Here's a test script to check:
WITH tblRentalUnit AS
(
SELECT 1 AS id, 100 AS MonthlyRent, 1 AS PropertyID
UNION ALL
SELECT 2 AS id, 300 AS MonthlyRent, 2 AS PropertyID
),
tblProperty AS
(
SELECT 1 AS id, 'Owner 1' AS OwnerName
UNION ALL
SELECT 2 AS id, 'Owner 2' AS OwnerName
),
tblTenant AS
(
SELECT 1 AS id, 1 AS UnitID
UNION ALL
SELECT 2 AS id, 1 AS UnitID
)
SELECT p.OwnerName, SUM(ru.MonthlyRent) AS PotentialRent, SUM(cnt) AS ExpectedRent
FROM tblRentalUnit ru
LEFT JOIN
tblProperty p
ON p.ID = ru.PropertyID
OUTER APPLY
(
SELECT COUNT(t.id) * ru.MonthlyRent AS cnt
FROM tblTenant t
WHERE t.UnitID = ru.ID
) td
GROUP BY p.OwnerName
What is the meaning of the sum of the unitMonthlyRent times the number of tenants, for some partiicular rental unit (COUNT(t.ID) * ru.MonthlyRent )?
Is it the case that all you are trying to do is see the difference between the total potential rent from all untis versus the expected rent (From only occcupied units) ? If so, then try this
Select p.OwnerName,
Sum(r.MonthlyRent) AS PotentinalRent,
Sum(Case t.Id When Null Then 0
Else r.MonthlyRent End) ExpectedRent
From tblRentalUnit r
Left Join tblTenant t
On t.UnitID = r.ID
left Join tblProperty p
On p.ID = r.PropertyID)
Group By p.OwnerName