Question

I’m trying to use an alias created in a SELECT in a WHERE statement. It doesn’t work and I read why in another SO question.

How do I make this work without repeating the subquery?

SELECT p.PatientID, p.PatientType, p.AccountNumber,
    p.FirstName + ' ' + p.LastName PatientFullName,
    p.CreatedDate, DATEDIFF(hour, p.CreatedDate, GETDATE()) TotalTime,
    (SELECT AVG(BGValue)
        FROM BloodGlucose
        WHERE PatientID = p.PatientID) AvgBG
FROM Patients p
WHERE AvgBG > 60;

This works:

SELECT p.PatientID, p.PatientType, p.AccountNumber,
    p.FirstName + ' ' + p.LastName PatientFullName, p.CreatedDate,
    DATEDIFF(hour, p.CreatedDate, GETDATE()) TotalTime,
    (SELECT AVG(BGValue) FROM BloodGlucose WHERE PatientID = p.PatientID) AvgBG
FROM Patients p
WHERE (SELECT AVG(BGValue) FROM BloodGlucose WHERE PatientID = p.PatientID) > 60;

But I don’t want to repeat that subquery. And I suspect it has poor performance.

Answer 1

Try using a derived table instead.

SELECT p.PatientID, p.PatientType, p.AccountNumber,
    p.FirstName + ' ' + p.LastName PatientFullName,
    p.CreatedDate, DATEDIFF(hour, p.CreatedDate, GETDATE()) TotalTime,
    bg.AvgBG 
FROM Patients p 
JOIN (SELECT PatientID, AVG(BGValue) AvgBG
    FROM BloodGlucose group by PatientID) BG
ON BG.PatientID = p.PatientID 
WHERE AvgBG > 60;

Derived tables work in sets and correlated subqueries work row-by-agonizing-row, that is why mine is faster.