6EQUJ5HD209458b
6EQUJ5HD209458b

Reputation: 301

PostgreSQL - Aliases column and HAVING

SELECT  
CASE WHEN SUM(X.Count)*3600 is null THEN  '0'  
            ELSE  
            SUM(X.Count)*3600  
       END AS PJZ,  
       X.Mass  
FROM X  
WHERE X.Mass > 2000  
HAVING ((X.Mass / PJZ * 100) - 100) >= 10;

Getting: ERROR: Column »pjz« doesn't exists.

How can I do something like this?

Upvotes: 24

Views: 17419

Answers (3)

Maciej
Maciej

Reputation: 10825

Other option is to surround query by WITH statement - for example:

WITH x as (
  SELECT coalesce(SUM(X.Count)*3600, 0) AS PJZ, X.Mass
  FROM X
  WHERE X.Mass > 2000
)
SELECT * from X WHERE PJZ >=10

It is far better then code duplication in my opinion

Upvotes: 11

user330315
user330315

Reputation:

Wrap it into a derived table:

SELECT CASE 
          WHEN PJZ = 0 THEN 100
          ELSE PJZ
       END as PJZ,
       mass
FROM (
    SELECT CASE 
             WHEN SUM(X.Count)*3600 is null THEN '0'  
             ELSE SUM(X.Count)*3600  
           END AS PJZ,  
           X.Mass  
    FROM X  
    WHERE X.Mass > 2000  
    GROUP BY X.mass
) t
WHERE PJZ = 0 
   OR ((X.Mass / PJZ * 100) - 100) >= 10;

(Note that I added the missing group by as otherwise the query would not be valid)

Upvotes: 7

jishi
jishi

Reputation: 24624

You can't use aliases in a having, and have to duplicate the statement in the having cluause. Since you only want to check for null, you could do this:

SELECT coalesce(SUM(X.Count)*3600, 0) AS PJZ, X.Mass
FROM X
WHERE X.Mass > 2000
HAVING ((X.Mass / coalesce(SUM(X.Count)*3600, 0) * 100) - 100) >= 10; 

Upvotes: 14

Related Questions