My CASE statement is wrong. Any idea what I am doing wrong? - sql

Question

I am in a logjam.
When I run the following query, it works:
select DISTINCT l.Seating_Capacity - (select count(*)
from tblTrainings t1, tbllocations l
where l.locationId = t1.LocationId) as
availableSeats
from tblTrainings t1, tbllocations l
where l.locationId = t1.LocationId
However, we would like to add a CASE statement that says, when Seating_Capacity - total count as shown above = 0 then show 'FULL' message.
Otherwise, show remaining number.
Here is that query:
select DISTINCT case when l.Seating_Capacity - (select count(*)
from tblTrainings t1, tbllocations l
where l.locationId = t1.LocationId) = 0 then 'full' else STR(Seating_Capacity) end)
availableSeats
from tblTrainings t1, tbllocations l
where l.locationId = t1.LocationId
I am getting 'Incorrect syntax near ')' which is near 'End'
I am also getting an error that the inner Seating_Capacity is invalid column name.
Your assistance is greatly appreciated.
I must have been in a dream land because I thought it was working during testing.
Now, the app is live and it isn't working.
Thanks a lot in advance
select DISTINCT l.LocationId,c.courseId, c.coursename, l.Seating_Capacity - (select count(*)
from tblTrainings t1
where l.locationId = t1.LocationId and c.courseId = t1.courseId) as
availableSeats,d.dateid,d.trainingDates,d.trainingtime,c.CourseDescription,
i.instructorName,l.location,l.seating_capacity
from tblLocations l
Inner Join tblCourses c on l.locationId = c.locationId
left join tblTrainings t on l.locationId = t.LocationId and c.courseId = t.courseId
Inner Join tblTrainingDates d on c.dateid=d.dateid
Inner Join tblCourseInstructor ic on c.courseId = ic.CourseId
Inner Join tblInstructors i on ic.instructorId = i.instructorId
WHERE CONVERT(VARCHAR(10), d.trainingDates, 101) >= CONVERT(VARCHAR(10), GETDATE(), 101)

Answer 1

To avoid repeating the expression, you can use a WITH clause to simplify your query:
WITH (
-- Start with your query that already works
SELECT DISTINCT l.Seating_Capacity - (select count(*)
from tblTrainings t1, tbllocations l
where l.locationId = t1.LocationId) AS availableSeats
FROM tblTrainings t1, tbllocations l
WHERE l.locationId = t1.LocationId
) AS source
SELECT
-- Add a CASE statement on top of it
CASE WHEN availableSeats = 0 THEN 'Full'
ELSE STR(availableSeats)
END AS availableSeats
FROM source

Answer 2

You have an extra ) at the end of your case statement remove that.
0 then 'full' else STR(Seating_Capacity) end)
^^^
for Seating_Capacity try accessing it with table alias like l.Seating_Capacity

Answer 3

I think you are over complicating the query with your subquery. As I understand it then the following should work as you need:
SELECT AvailableSeats = CASE WHEN l.Seating_Capacity - COUNT(*) = 0 THEN 'Full'
ELSE STR(l.Seating_Capacity - COUNT(*))
END
FROM tblTrainings t1
INNER JOIN tblLocations l
ON l.LocationID = t1.LocationID
GROUP BY l.Seating_Capacity;
I have changed your else to STR(l.Seating_Capacity - COUNT(*)) because I assume you want to know the seats remaining, rather than just the capacity? If I have misinterpreted the requirement, just change it to STR(l.Seating_Capacity).
I have also switched your ANSI 89 implicit joins to ANSI 92 explicit joins, the standard changed over 20 years, and there are plenty of good reasons to switch to the newer syntax. But for completeness the ANSI 89 version of the above query would be:
SELECT AvailableSeats = CASE WHEN l.Seating_Capacity - COUNT(*) = 0 THEN 'Full'
ELSE STR(l.Seating_Capacity - COUNT(*))
END
FROM tblTrainings t1, tblLocations l
WHERE l.LocationID = t1.LocationID
GROUP BY l.Seating_Capacity;
EDIT
To adapt your full query you can simply replace your subquery in the select, with a joined subquery:
SELECT l.LocationId,
c.courseId,
c.coursename,
CASE WHEN l.Seating_Capacity - t.SeatsTaken = 0 THEN 'Full'
ELSE STR(l.Seating_Capacity - t.SeatsTaken)
END AS availableSeats,
d.dateid,
d.trainingDates,
d.trainingtime,
c.CourseDescription,
i.instructorName,
l.location,
l.seating_capacity
FROM tblLocations l
INNER JOIN tblCourses c
ON l.locationId = c.locationId
LEFT JOIN
( SELECT t.LocationID, t.CourseID, SeatsTaken = COUNT(*)
FROM tblTrainings t
GROUP BY t.LocationID, t.CourseID
) t
ON l.locationId = t.LocationId
AND c.courseId = t.courseId
INNER JOIN tblTrainingDates d
ON c.dateid=d.dateid
INNER JOIN tblCourseInstructor ic
ON c.courseId = ic.CourseId
INNER JOIN tblInstructors i
ON ic.instructorId = i.instructorId
WHERE d.trainingDates >= CAST(GETDATE() AS DATE);
JOINs tend to optimise better than correlated subqueries (although sometimes the optimiser can determine that a JOIN would work instead), it also means that you can reference the result (SeatsTaken) multiple times without re-evaluating the subquery.
In addition, by moving the count to a subquery, and removing the join to tblTrainings I think you eliminate the need for DISTINCT which should improve the performance.
Finally I have changed this line:
WHERE CONVERT(VARCHAR(10), d.trainingDates, 101) >= CONVERT(VARCHAR(10), GETDATE(), 101)
To
WHERE d.trainingDates >= CAST(GETDATE() AS DATE);
I don't know if you do, but if you had an index on d.TrainingDates then by converting it to varchar to compare it to today you remove the ability for the optimiser to use this index, since you are only saying >= midnight today, there is no need to perform any conversion on d.TrainingDates, and all you need to do is remove the time portion of GETDATE(), which can be done by casting to DATE. More on this is contained in this article (Yet another gem from Aaron Bertrand)