Columns in the SELECT not in the GROUP BY
I have an SQL query where I am using the SUM function with the GROUP BY clause. I need to select a couple more columns from the table without adding those columns in the GROUP BY clause.
The answer is yes, but only in a very certain way.
Let's imagine that this is your current query:
select foo , sum(bar) as sumbar from qux group by foo
Now, we want to add two more columns—let's call them other1 and other2—but we don't want to add them to the GROUP BY, because we still want only one row per foo in the results.
The very certain and only way that this will be allowed is by ensuring that the two new columns are inside aggregate functions.
Here's one example of many possibilities:
select foo , sum(bar) as sumbar , max(other1) as maxcol1 , avg(other2) as avgcol2 from qux group by foo
What is NOT ALLOWED is to put "naked" columns (not inside aggregate functions) into the SELECT clause but leave them out of the GROUP BY clause:
select foo , sum(bar) as sumbar , other1 , other2 from qux group by foo
This is invalid syntax, and only MySQL will run it (although they do warn you, in GROUP BY and HAVING with Hidden Fields, that results can be unpredictable).
Dig Deeper on Oracle development languages
Related Q&A from Rudy Limeback
Using the SQL GROUP BY clause for counting combinations
Read SQL expert Rudy Limeback's advice for counting combinations in a table with SQL's GROUP BY clause Continue Reading
How to sort an SQL UNION query with special ORDER BY sequence
SQL expert Rudy Limeback explains how to sort an SQL UNION query using a special ORDERY BY sequence. Continue Reading
Using an SQL SELECT statement from a non-existing table
SQL expert Rudy Limeback explains how to formulate a query using an SQL SELECT statement from a non-existing table. Continue Reading