Using DATEDIFF within a SELECT - sql

Question

SELECT DISTINCT
Test,
FIRST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) StartDate,
LAST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) EndDate,
DATEDIFF(MONTH, StartDate, EndDate)
FROM
Tests
Any ideas on how I can include the difference between the 2 dates within the SELECT?
I want to be able to use the values StartDate and EndDate however I'm getting an error saying StartDate and EndDate are invalid columns.

Answer 1

You can't use aliases you created in the SELECT clause elsewhere in the same SELECT clause, so you need to place your FIRST_VALUE() and LAST_VALUE() functions within the DATEDIFF() function:
SELECT DISTINCT
Test,
FIRST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) StartDate,
LAST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) EndDate,
DATEDIFF(MONTH, FIRST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC)
, LAST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) )
FROM
Tests
Or you could run everything but the DATEDIFF() in a subquery/cte:
SELECT *,DATEDIFF(MONTH, StartDate, EndDate)
FROM (SELECT DISTINCT
Test,
FIRST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) StartDate,
LAST_VALUE (CONVERT(DATE, datetime)) OVER (ORDER BY datetime DESC) EndDate
FROM
Tests
) AS sub
You also can't use aliases created in the SELECT clause in the WHERE clause, but you can use them in ORDER BY and of course in any outer queries.

Answer 2

As a note, I think your query is equivalent to:
SELECT Test,
CONVERT(DATE, MIN(datetime)) as StartDate,
CONVERT(DATE, MAX(datetime) as EndDate,
DATEDIFF(MONTH, CONVERT(DATE, MIN(datetime)), CONVERT(DATE, MAX(datetime)))
FROM Tests
GROUP BY Test;
The functions MIN() and MAX() are more succinct. In addition, the query is easier to read as a group by rather than as a select distinct.