Question

I have three columns in a table – ID, Column1, Column2 – with this example data:

ID  Column1 Column2
----------------------
1   1       2  
2   2       1  
3   4       3  
4   3       4

Since, in the first two rows, Column1 and Column2 have the same values (but in different columns), I want my MAX query to return an ID of 2. Same thing with rows 3 and 4 …. since Columns 1 and 2 have the same values (but in different columns), I want MAX(ID) to return 4. Of course, with MAX, you use Group By, but that will not work in my case.

In effect, I need a Group By to work across two columns. Is this possible? If not, what’s the best way to accomplish getting the IDs of 2 and 4 given the matching values that are in different columns?

Answer 1

An alternative to use CASE would be to UNION the results of the column(s), and perform a MAX on them:

  SELECT z.id,
         MAX(z.col)
    FROM (SELECT x.id,
                 x.column1 AS col
            FROM TABLE x
          UNION 
          SELECT y.id,
                 y.column2
            FROM TABLE y) z
GROUP BY z.id