Count distinct values on sevaral fields 1

[graabein]

Hi, I have a sort of stupid table structure on the project I inherited. It's a list of what work was performed that uses codes like this:

CustomerID | Date | Code1 | Code2 | Code3 | Code4

I want to know all work that was performed for a certain period and thus need to count the distinct codes on codes 1-4.

Say I have the test data:

1 | today | AA | AB | AC | Null
2 | today | AC | BB | Null | Null

I want to get the following:

AA | 1
AB | 1
AC | 2
BB | 1

The table is pretty huge so I need an effective query... Any hints?


graabein

 

[r937]

SELECT daCode
     , COUNT(*) AS daCount
  FROM ( SELECT Code1 AS daCode
           FROM daTable
          WHERE Code1 IS NOT NULL
         UNION ALL
         SELECT Code2 
           FROM daTable
          WHERE Code2 IS NOT NULL
         UNION ALL
         SELECT Code3 
           FROM daTable
          WHERE Code3 IS NOT NULL
         UNION ALL
         SELECT Code4 
           FROM daTable
          WHERE Code4 IS NOT NULL
       ) AS d
GROUP
    BY daCode

r937.com | rudy.ca
Buy my new book Simply SQL from Amazon

 

[graabein]

I think I did it! Check this out:

select Work.Code, count(*) from (
select Code1 as Code from table1
union all
select Code2 as Code from table1
) as Work
where Work.Code is not null
group by Work.Code

What do you think?


graabein

 

[graabein]

LOL nice coincidence! Thanks man.


graabein