Distinct operation using GROUP BY

Question

I was doing some research on retrieving Duplicate records in a table.
create table x
(empid number,
empname varchar2(50)
insert into x values (1,'John');
insert into x values (2,'Reynolds');
insert into x values (3,'Harrison');
insert into x values (1,'Kate');
insert into x values (2,'Hans');
SQL> select * from x;
     EMPID EMPNAME
         1 John
         2 Reynolds
         3 Harrison
         1 Kate
         2 Hans
SQL> select empid from x group by empid;          ------ Query 1
     EMPID
         1
         2
         3        
SQL> select min(empid) from x group by empid;      ------ Query 2
MIN(EMPID)
         1
         2
         3
SQL> select * from x where empid not in
  2  (select min(empid) from x group by empid);   -------- Query 3   
no rows selectedI understand how Query 1 and Query 2 retrieves distinct values. Query 3 can't fetch duplicate values because it is equivalent to
SQL> select * from x where empid not in
  2  (select distinct(empid) from x);
no rows selectedAm i right?
From various OTN threads, i gathered that a query using ROWID should be used for retrieving duplicate rows.
Something like
select * from x where rowid not in
(select min(rowid) from x group by empid);   ------------- Query 4How come this query (Query 4) works and Query 3 doesn't?

Answer 1

Hi,
Y.Ramlet wrote:
I was doing some research on retrieving Duplicate records in a table.
create table x
(empid number,
empname varchar2(50)
insert into x values (1,'John');
insert into x values (2,'Reynolds');
insert into x values (3,'Harrison');
insert into x values (1,'Kate');
insert into x values (2,'Hans');
SQL> select * from x;
EMPID EMPNAME
1 John
2 Reynolds
3 Harrison
1 Kate
2 Hans
SQL> select empid from x group by empid;          ------ Query 1
EMPID
1
2
3        
SQL> select min(empid) from x group by empid;      ------ Query 2
MIN(EMPID)
1
2
3
SQL> select * from x where empid not in
2  (select min(empid) from x group by empid);   -------- Query 3   
no rows selectedI understand how Query 1 and Query 2 retrieves distinct values. Query 3 can't fetch duplicate values because it is equivalent to
SQL> select * from x where empid not in
2  (select distinct(empid) from x);
no rows selectedAm i right?Yes, you're right. Query 3 and Query 4 get the same results (or lack of results) by different ways.
From various OTN threads, i gathered that a query using ROWID should be used for retrieving duplicate rows.
Something like
select * from x where rowid not in
(select min(rowid) from x group by empid);   ------------- Query 4How come this query (Query 4) works and Query 3 doesn't?They both work; they get different results because they are doing different things.
ROWID is unique; empid is not. When you "GROUP BY empid", any group with more than one member will have ROWIDs that are not equal to MIN (ROWID) for that group.