Has anyone encountered this problem ?

Question

hi ,
i have used the tranpose funtion that uses stringaggtype written by either Scott or tom Ktye
my data is as follows :
ID     OP     RD     IDS     M_SI     Ma_VAL     Mi_VAL     idx     NVALUE
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     1     2868.44
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     2     2836.32
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     3     2836.05
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     4     2816.53
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     5     2829.23
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     6     2789.23
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     7     2813.42
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     8     2825.63
1123r     gfrr     10/10/2006 19:15     1124t     9     2868.44     2789.23     9     2798.73i did a tranpose using :
select a.po , a.rd , a.ids ,           
tranposefn(a.nvalue)  siteval          
from bt_111 a          
group by a.id, a.op , a.rd , a.pids           
order by  a.id, a.op , a.rd , a.pids           and results :
gfrr     10/10/2006 19:15     1124t     2868.44,2836.32,2836.05,2829.23,2813.42,2798.73,2825.63,2789.23,2816.53notice that the 4th value 2816.53 becomes the last after transposing
and i tried to transpose by each row until the last row i.e 9
test1 :
select a.po , a.rd , a.ids ,
tranposefn(a.nvalue)  siteval
from bt_111 a
where rownum <= 1 - 9
group by a.id, a.op , a.rd , a.pids
order by  a.id, a.op , a.rd , a.pids from test1 when i did it for rownum <= 4 the results is still ok
but when i did it for rownum <=5 , my 4th values will become the fifth and that 4th value keeps becoming the last value afterwards
i do not believe it's something to do with the transpose function but how does Oracle order the records ??
pls help
tks & rdgs

Answer 1

Possibly it may work for simplistic GROUP BY. See counter case however...
Oracle Database 10g Enterprise Edition Release 10.2.0.2.0 - Production
With the Partitioning, OLAP and Data Mining options
SQL> SELECT CONCAT_ALL (CONCAT_EXPR (productnumber, ',')) aggregate
  2  FROM  (SELECT DISTINCT SUBSTR (productnumber, 1, 2) productnumber
  3         FROM   prdspc
  4         WHERE  productnumber NOT LIKE 'FB%'
  5         ORDER BY SUBSTR (productnumber, 1, 2));
AGGREGATE
CO,ED,ER,HA,LG,LO,MO,NO,PA,PC,PR,SA,SI,VI,VK
SQL> SELECT productspctypeid,
  2         CONCAT_ALL (CONCAT_EXPR (productnumber, ',')) aggregate
  3  FROM  (SELECT DISTINCT productspctypeid,
  4                SUBSTR (productnumber, 1, 2) productnumber
  5         FROM   prdspc
  6         WHERE  productnumber NOT LIKE 'FB%'
  7         ORDER BY SUBSTR (productnumber, 1, 2))
  8  GROUP BY productspctypeid;
PRODUCTSPCTYPEID AGGREGATE
               2 ER,LG,LO,NO,SA,SI,VK,MO
            3069 VI
            3120 HA
            3292 PC,VI
            3505 ER,LO,VK,SI,SA,NO,MO,LG
            3523 ER,LG,NO,SA,MO,LO,SI,VK
            4011 PA
            4012 PR
            4013 PA
            4014 PA
            4043 PA
            4047 ED
            4048 CO
13 rows selected.
SQL>