Need help changing query

Question

There has been change to requirements and i'm not sure how to change my query to handle the changes.
The table basically has entries for each end point of a connection and the query combines the from/to row. Now the requirement change is that the connection can have multiple end points with a single "from" end point. SO the result set may be multiple rows for a connection with the same "from endpoint" and the different "to endpoints".
The query is below ( Sorry about the bad formatting)
Basically the "seq<=2" is no longer true and there is no finite limit. If there are multiple endpoints, the key is that for each result "row" the "from endpoint" will be same but different "to endpoints"
I can do a "SIGN(seq-1),1 " to get the multiple rows but don't know how to remove the "max" without getting a "not a GROUP BY expression" error.
Please help. Thank you.
SELECT tt.descr TYPE,
       from_id,
       from_name,
       from_ra,
       from_field1,
       from_ran,
       total,
       to_id,
       to_name,
       to_ra,
       to_field2,
       to_ran
FROM   (SELECT   TYPE,
                 Max(Decode(seq,1,id,
                                NULL)) from_id,
                 Max(Decode(seq,1,NAME,
                                NULL)) from_name,
                 Max(Decode(seq,1,ra,
                                NULL)) from_ra,
                 Max(Decode(seq,1,field1,
                                NULL)) from_field1,
                 Max(Decode(seq,1,ran,
                                NULL)) from_ran,
                 Max(Decode(seq,1,total,
                                NULL)) total,
                 Max(Decode(seq,2,id,
                                NULL)) to_id,
                 Max(Decode(seq,2,NAME,
                                NULL)) to_name,
                 Max(Decode(seq,2,ra,
                                NULL)) to_ra,
                 Max(Decode(seq,2,field1,
                                NULL)) to_field2,
                 Max(Decode(seq,2,ran,
                                NULL)) to_ran
        FROM     (SELECT tablew.group                                                     AS group,
                         tablew.id                                                        AS id,
                         s.NAME,
                         Nvl((SELECT t.field1
                              FROM   tableb t
                              WHERE  t.NAME = (SELECT NAME
                                               FROM   tablea
                                               WHERE  id = another_id)
                                     AND t.field2 = tablew.ra
                                     AND ROWNUM <= 1),'No Data') AS field1,
                         tablew.RANGE,
                         tablew.total                                                     AS total,
                         tablew.ra,
                         tablew.TYPE,
                         Row_number()
                           OVER(PARTITION BY tablew.group ORDER BY tablew.cla, tablew.id) seq
                  FROM   tablew w,
                         tables s
                  WHERE  tablew.id LIKE '"+var1+"%'
                         AND s.id = tablew.id)
        WHERE    seq <= 2
        GROUP BY group,
                 TYPE) c,
       tablec tt
WHERE  c.TYPE = tt.TYPE;

Answer 1

Hi,
Welcome to the forum!
Sorry, I'm not familiar with your tables or yopur application.
Can you post a few rows of sample data, and the results you want from that data? For the sample data, executable CREATE TABLE and INSERT statements (or CREAT TABLE AS statements) are best.
Simplify as much as you can. for example, it looks like you're currently producing two sets of five columns (id, name, ra field and ran). Is it practical to ignore, say, name, ra and field for now? Once you learn how to handle id and ran, you'll probably know how to handle the other columns as well.
What version of Oracle are you using? (Always post this when you ask a question. You don't want to try an answer that won't work any more than I want to write an answer that won't work for you.)
user11033180 wrote:
The query is below ( Sorry about the bad formatting)The query is actually formatted much better than average!