Reputation: 279
I have the query below
SELECT distinct
count(patient.id)
CAST(YEAR(completiondate) AS VARCHAR(4)) + '-' + CAST(MONTH(completiondate) AS VARCHAR(2)) AS Mjesec
FROM ....
WHERE
incident.completionDate Between Convert(smalldatetime, '01/10/2007', 103) and Convert(smalldatetime, '01/11/2013', 103)
and servicecharges.serviceid in (47)
and incident.status != 6
and ServiceRequestDescription LIKE 'something'
and chargeDescr = 'test'
group by CAST(YEAR(completiondate) AS VARCHAR(4)) + '-' + CAST(MONTH(completiondate) AS VARCHAR(2))
the result i take, have unique rows of patient.id AND Year-Month. That means that i can have multiple rows with the same patient.id and several year-month.
how can i change that query, in order to take the results, with each patient.id only once, with the first completiondate with the specific elements?
Upvotes: 1
Views: 190
Reputation: 1106
Please Try it
WITH CTE AS
(
SELECT *,RN=ROW_NUMBER() OVER (PARTITION BY patient ORDER BY patient DESC) FROM tablename
)
select * from CTE where RN>1
Upvotes: 0
Reputation: 24144
I guess it should be:
SELECT
patient.id
MIN(
CAST(YEAR(completiondate) AS VARCHAR(4)) + '-' + CAST(MONTH(completiondate) AS VARCHAR(2))
)
AS Mjesec
FROM ....
WHERE
incident.completionDate Between Convert(smalldatetime, '01/10/2007', 103) and Convert(smalldatetime, '01/11/2013', 103)
and servicecharges.serviceid in (47)
and incident.status != 6
and ServiceRequestDescription LIKE 'something'
and chargeDescr = 'test'
group by patient.id
Upvotes: 3