can you tell me scèneario.

The scenario's.. i want, duplicate records from a table to a
target.. n remaining records to b on another target...

Hi,
If I understand your question correctly, then it is like- 
How to send original (distinct) records to one target 
(let's say TA) and duplicates to another target (let's say 
TB).
It can simply be done as follows - 
In your mapping, take the source table containing whole 
records (duplicates as well). Now, take 2 source qualifier 
(SQ) transformations. The first SQ should 
contain 'Distinct' sql query (hope you know that) and the 
output ports should go to TA table.
The second SQ should contain the following sql stmt (which 
filters only the duplicate records) - 
-------------------------------------------
SELECT empno, ename, job, mgr, hiredate, sal, comm, deptno
FROM (SELECT empno, ename, job, mgr, hiredate, sal, comm, 
deptno, ROW_NUMBER() OVER (PARTITION BY empno ORDER BY 
EMPNO) rn FROM s_assign_1_emp)
WHERE rn > 1
-------------------------------------------
I am using emp table example (hope you can understand it).
Now, the output ports of the seecond SQ should go to TB 
table.
This is the simplest way of doing it. Hope it will help.

If you feel any issue, please let me know at 
shrikant_rps@yahoo.com

Hi You can do this another way also
put the expresession transforamtion and read some previous
record with the currect record or any unique column value in
the expression and if they are equal then assign the
flag(means duplicate flag) else
assign the new flage.
pass the duplicate flag records to the filter and goes to
the TB(duplicate records target) and unique values to the
origonal target

Hi,
Yes you can do it by using Expression transformation as 
well. Infact, by using 2 SQ (source qualifier) and using 
sql queries (as I explained earlier), you can perform the 
required task on relational tables only (but not on flat 
files etc).
In the case of flat files as source, you will need to use 
Expression transformation. But you also need to use sorter 
before expression (otherwise record flag will change for 
every incoming record & duplicates won't be traced). The 
sequence of transformations will be as follows - 

Source -> SQ -> Sorter ->Expression ->Router(or 2 filters) -
>Targets