Try this:

SELECT COUNT(*) FROM
(
SELECT custid, edate
FROM  dbw_promoc
MINUS
SELECT custid, edate
FROM promo@report
)
;

Clive Moore schrieb:

SELECT d.document_id
FROM document d, courier_info ci
WHERE d.document_id = ci.doc_inv_id
ORDER BY ci.order_field

courier_info table. <<

Rows are not records; fields are not columns; tables are not files.  A
result set has be turned into a CURSOR (a sequential file structure)
before it can be sorted.  Since the cursor is created after the result
set, it can only use column names in the SELECT list.  YOur sort
column has to be in your SELECT list.

I think that you meant to use plural or collective nouns for table
names; I am assuming that you have more than one document.  The name
"courier_info" is weird; how would a table about couriers not contain
information about couriers?  This is like putting "tbl-" prefixes on a
table name -- what else would be?

Since you did not bother to post DDL or even tell us the name of the
sort column, let me make wild guessses:

DECLARE Foobar CURSOR FOR
SELECT D1.document_id, C1.sort_col
  FROM Documents AS D1, Couriers AS C1
 WHERE D1.document_id = C1.doc_inv_id
 ORDER BY sort_col;

Of course, if there is a 1:m relationship between Documents and
Couriers, you will get multiple rows.  Since the multiple sort_col
values can vary within each document_id, there is no way to get just
the document_id from what you posted.  We have to pick one value and
assume that document_id is unique:

DECLARE Foobar CURSOR FOR
SELECT D1.document_id, MIN(C1.sort_col) AS min_sort_col
  FROM Documents AS D1, Couriers AS C1
 WHERE D1.document_id = C1.doc_inv_id
 GROUP BY D1.document_id
 ORDER BY min_sort_col;