Did I misuse 'DISTINCT' in this SQL statement?

Question

I have a table TI_ORDER only contains following data:
O_ID E_CODE M_SEQ B_SEQ
VARCHAR NUMBER NUMBER NUMBER
CE013 1 1 1
CE013 1 2 1
CE013 1 3 1
CE013 1 4 1
CE013 1 5 1
CE013 1 6 1
CE013 1 6 2
CE013 1 7 1
CE013 1 8 1
CE013 1 8 2
CE013 1 9 1
CE013 1 10 1
CE013 1 10 2
CE013 2 1 1
CE013 2 2 1
CE013 2 3 1
CE013 2 4 1
CE013 2 5 1
CE013 2 6 1
CE013 2 6 2
CE013 2 7 1
CE013 2 8 1
CE013 2 8 2
CE013 2 9 1
CE013 2 10 1
CE013 2 10 2
If I execute this SQL:
==============================================
SELECT a.o_id, a.e_code,
COUNT(a.o_id) OVER (PARTITION BY a.o_id, a.e_code) AS cnt
FROM ( SELECT DISTINCT o_id, e_code, m_seq FROM ti_order ) a
WHERE a.o_id = 'CE013'
==============================================
It will show:
==============================================
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 1 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
CE013 2 10
=============================================
If I add 'DISTINCT' to previous SQL statement:
============================================
SELECT DISTINCT a.o_id, a.e_code,
COUNT(a.o_id) OVER (PARTITION BY a.o_id, a.e_code) AS cnt
FROM ( SELECT DISTINCT o_id, e_code, m_seq FROM ti_order ) a
WHERE a.o_id = 'CE013'
============================================
It displays:
============================================
CE013 1 13
CE013 2 13
============================================
Why does it not show following output as I want ?
============================================
CE013 1 10
CE013 2 10
============================================

Answer 1

Looks like you have stumbled across a bug. Below output indicates that 9i (example 2) gets the correct answer here while 8i (example 1) does not. You ARE on 8i right?
In any case, there is no need to use analytic functions here, good old COUNT is fine to get the correct answer (example 3).
--------------------------- example 1 -----------------------------
Oracle8i Enterprise Edition Release 8.1.7.4.0 - Production
With the Partitioning option
JServer Release 8.1.7.4.0 - Production
SQL> CREATE TABLE table_name (
2 o_id VARCHAR2 (5),
3 e_code NUMBER (2),
4 m_seq NUMBER (2),
5 b_seq NUMBER (2));
Table created.
SQL> INSERT INTO TABLE_NAME VALUES ('CE013','1','1','1');
1 row created.
SQL> INSERT INTO TABLE_NAME VALUES ('CE013','1','2','1');
(snip)
SQL> INSERT INTO TABLE_NAME VALUES ('CE013','2','10','2');
1 row created.
SQL> SELECT DISTINCT a.o_id, a.e_code,
2 COUNT (a.o_id) OVER (PARTITION BY a.o_id, a.e_code) AS cnt
3 FROM (SELECT DISTINCT o_id, e_code, m_seq
4 FROM table_name) a
5 WHERE a.o_id = 'CE013';
O_ID E_CODE CNT
CE013 1 13
CE013 2 13
SQL>
--------------------------- example 2 -----------------------------
Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production
With the Partitioning, OLAP and Oracle Data Mining options
JServer Release 9.2.0.4.0 - Production
SQL> CREATE TABLE table_name (
2 o_id VARCHAR2 (5),
3 e_code NUMBER (2),
4 m_seq NUMBER (2),
5 b_seq NUMBER (2));
SQL> INSERT INTO TABLE_NAME VALUES ('CE013','1','1','1');
1 row created.
SQL> INSERT INTO TABLE_NAME VALUES ('CE013','1','2','1');
1 row created.
(snip)
SQL> INSERT INTO TABLE_NAME VALUES ('CE013','2','10','2');
1 row created.
SQL> SELECT DISTINCT a.o_id, a.e_code,
2 COUNT (a.o_id) OVER (PARTITION BY a.o_id, a.e_code) AS cnt
3 FROM (SELECT DISTINCT o_id, e_code, m_seq
4 FROM table_name) a
5 WHERE a.o_id = 'CE013';
O_ID E_CODE CNT
CE013 1 10
CE013 2 10
SQL>
--------------------------- example 3 -----------------------------
Oracle8i Enterprise Edition Release 8.1.7.4.0 - Production
With the Partitioning option
JServer Release 8.1.7.4.0 - Production
SQL> SELECT a.o_id, a.e_code, COUNT (*)
2 FROM (SELECT DISTINCT o_id, e_code, m_seq
3 FROM table_name) a
4 WHERE a.o_id = 'CE013'
5 GROUP BY a.o_id, a.e_code;
O_ID E_CODE COUNT(*)
CE013 1 10
CE013 2 10
SQL>
Padders