Problem with query in cursor (Oracle 8i)

Question

Hi all,
Following query works perfectly fine if I test it in my development environment
SELECT dt,
      (SELECT 'Y'
         FROM dual
        WHERE EXISTS (SELECT 'x'
                        FROM top_days t
                       WHERE t.date = dt))
FROM (SELECT to_date(p_date_start, 'dd-mm-yyyy') + LEVEL - 1 dt
                   FROM dual
                 CONNECT BY LEVEL <=
                            to_date(p_date_end, 'dd-mm-yyyy') - to_date(p_date_start, 'dd-mm-yyyy') + 1
                        AND PRIOR dbms_random.value IS NOT NULL) --workaround for oracle 8i
                top_days t
WHERE dt = t.date(+)
ORDER BY dt;However, when I try to include it in my package (where it's used in a cursor) I get a PLS-00103: Encountered the symbol "SELECT" when expecting one of the following: ( - + mod null others <an identifier> etc)
CURSOR c_holidays IS
           SELECT dt,
                (SELECT 'Y'
                   FROM dual
                  WHERE EXISTS (SELECT 'x'
                           FROM top_days t
                          WHERE t.date = dt))
           FROM (SELECT to_date(p_date_start, 'dd-mm-yyyy') + LEVEL - 1 dt
                   FROM dual
                 CONNECT BY LEVEL <=
                            to_date(p_date_end, 'dd-mm-yyyy') - to_date(p_date_start, 'dd-mm-yyyy') + 1
                        AND PRIOR dbms_random.value IS NOT NULL) --workaround for oracle 8i
                top_days t
          WHERE dt = t.date(+)
          ORDER BY dt;Is there some special thing I don't know about Oracle 8i that prevents me from using this query in a cursor? Or am I just missing something blatantly obvious?
Note: p_date_start and p_date_end are parameters that are filled in the function
FYI: Oracle 8i release 8.1.7.4.1 and PL/SQL Release 8.1.7.4.0

Answer 1

Hi,
You can do a join instead of using a scalar sub-query:
SELECT        a.dt
,       CASE
          WHEN  t.dt IS NOT NULL
          THEN  'Y'
       END          AS in_top_days
FROM        (
          SELECT      to_date (p_date_start, 'dd-mm-yyyy') + LEVEL - 1 dt
                FROM dual
                CONNECT BY      LEVEL <=
                                 to_date(p_date_end, 'dd-mm-yyyy') - to_date(p_date_start, 'dd-mm-yyyy') + 1
                        AND      PRIOR dbms_random.value IS NOT NULL --workaround for oracle 8i
       )          a
,         top_days     t
WHERE        a.dt      = t.dt (+)
ORDER BY  a.dt; 
I hope this answers your question.
If not, post a little sample data (CREATE TABLE and INSERT statements, relevant columns only), a few sets of parameters (p_date_start and p_date_end), and the results you want from that data for each set.
Explain, using specific examples, how you get those results from that data.