MS-Access -> SELECT AS + ORDER BY = error

Question

I'm trying to make a query to retrieve the region which got the most sales for sweet products. 'grupo_produto' is the product type, and 'regiao' is the region. So I got this query:

SELECT TOP 1 r.nm_regiao,  (SELECT COUNT(*)
        FROM Dw_Empresa
        WHERE grupo_produto='1' AND 
        cod_regiao = d.cod_regiao) as total 
FROM Dw_Empresa d
INNER JOIN tb_regiao r ON r.cod_regiao = d.cod_regiao ORDER BY total DESC

Then when i run the query, MS-Access asks for the "total" parameter. Why it doesn't consider the newly created 'column' I made in the select clause?

Thanks in advance!

Conscience

Answer 1

Aliases are only usable in the query output. You can't use them in other parts of the query. Unfortunately, you'll have to copy and paste the entire subquery to make it work.

Corell

You can't use the alias, but you can repeat the expression being aliased. Jet will calculate it only once, so it's only inefficient in regard to typing. – Bit

Not in this case. You will get an error if you try to order by the posted subquery. – Ref

"Aliases are only useful in the query output" -- I disagree: I think it is very useful to be able to use 'aliases' in the ORDER BY clause. – Interstate

"You can't use ['aliases'] in other parts of the query" -- that's because those 'other parts' are resolved before the SELECT clause. However, the ORDER BY clause is applied after the SELECT clause so it shouldn't be a problem... it shouldn't be but in Access it is :( – Interstate

Aliases are useful throughout the entire query, including other fields in the SELECT list. I should have said "usable." I will edit to clarify. – Corell

I haven't delved into the SQL enough to tell, but I'm not certain it requires a subquery. Seems to me it might be doable with a JOIN on a derived table (i.e., in the FROM clause). But as I say, I'm not looking that closely at the SQL. – Bit

Answer 2

Old Question I know, but it may help someone knowing than while you cant order by aliases, you can order by column index. For example, this will work without error :

SELECT 
 firstColumn,
 IIF(secondColumn = '', thirdColumn, secondColumn) As yourAlias
FROM
 yourTable
ORDER BY
 2 ASC

The results would then be ordered by the values found in the second column wich is the Alias "yourAlias".

Humpage

Answer 3

You can do it like this

select * from(
  select a + b as c, * from table)
  order by c

Access has some differences compared to Sql Server.

Rabush

Answer 4

Why it doesn't consider the newly created 'column' I made in the select clause?

Because Access (ACE/Jet) is not compliant with the SQL-92 Standard.

Consider this example, which is valid SQL-92:

SELECT a AS x, c - b AS y
  FROM MyTable
 ORDER
    BY x, y;

In fact, x and y the only valid elements in the ORDER BY clause because all others are out of scope (ordinal numbers of columns in the SELECT clause are valid though their use id deprecated).

However, Access chokes on the above syntax. The equivalent Access syntax is this:

SELECT a AS x, c - b AS y
  FROM MyTable
 ORDER
    BY a, c - b;

However, I understand from @Remou's comments that a subquery in the ORDER BY clause is invalid in Access.

Interstate

Answer 5

Try using a subquery and order the results in an outer query.

SELECT TOP 1 * FROM
(
    SELECT
        r.nm_regiao, 
        (SELECT COUNT(*)
         FROM Dw_Empresa
         WHERE grupo_produto='1' AND cod_regiao = d.cod_regiao) as total 
    FROM Dw_Empresa d
    INNER JOIN tb_regiao r ON r.cod_regiao = d.cod_regiao
) T1
ORDER BY total DESC

(Not tested.)

Inapplicable

It's weird. When I try that, the "total" column becomes all 0 and instead of one row showing up, I get 10 similar rows. I'm trying to figure out why. That's the way, though. Thanks – Conscience

Answer 6

How about:

SELECT TOP 1  r.nm_regiao 
FROM (SELECT Dw_Empresa.cod_regiao, 
             Count(Dw_Empresa.cod_regiao) AS CountOfcod_regiao
      FROM Dw_Empresa
      WHERE Dw_Empresa.[grupo_produto]='1'
      GROUP BY Dw_Empresa.cod_regiao
      ORDER BY Count(Dw_Empresa.cod_regiao) DESC) d
INNER JOIN tb_regiao AS r 
ON d.cod_regiao = r.cod_regiao

Ref

Answer 7

I suggest using an intermediate query.

 SELECT r.nm_regiao, d.grupo_produto, COUNT(*) AS total
   FROM Dw_Empresa d INNER JOIN tb_regiao r ON r.cod_regiao = d.cod_regiao
   GROUP BY r.nm_regiao, d.grupo_produto;

If you call that GroupTotalsByRegion, you can then do:

SELECT TOP 1 nm_regiao, total FROM GroupTotalsByRegion 
  WHERE grupo_produto = '1' ORDER BY total DESC

You may think it's extra work to create the intermediate query (and, in a sense, it is), but you will also find that many of your other queries will be based off of GroupTotalsByRegion. You want to avoid repeating that logic in many other queries. By keeping it in one view, you provide a simplified route to answering many other questions.

Tifanie

You don't have to save it -- you should be able to use a derived table in the FROM clause in place of the saved QueryDef. – Bit

Answer 8

How about use: WITH xx AS ( SELECT TOP 1 r.nm_regiao, (SELECT COUNT(*) FROM Dw_Empresa WHERE grupo_produto='1' AND cod_regiao = d.cod_regiao) as total FROM Dw_Empresa d INNER JOIN tb_regiao r ON r.cod_regiao = d.cod_regiao ) SELECT * FROM xx ORDER BY total

Mercedes