ODI11G Union All

Question

Hi,
I am trying to implement Union All in ODI11g. I have a plsql procedure and want to re-implement it in ODI11G.
Mine procedure is like:
INSERT INTO STAGE_TABLE1
(CustomerType, other fields)
select 'RETAILER' AS customertype,other fields
from table1 ,table2, table3
where table1.col1= table2.col2
and table2.col2 = table3.col3
UNION ALL
SELECT 'SUPERMARKET' AS CUSTOMERTYPE, OTHER FIELDS
FROM TABLE4, TABLE5, TABLE6
WHERE TABLE4.COL1 = TABLE5.COL7
AND TABLE6.COL8 = TABLE5.COL3
For applying Union, I add two datasets in the interface. Dataset1 is the first part of the insert script and dataset2 is the second part of the insert script.
In the select scripts, customer type retailer and supermarket are hardcoded literals , they are not a column of a table.
So I took the error message below when I want to execute the interface:
I am having the error message below:
"Target Mapping for Column CustomerType in DataSet DATASET1     A mapping executed on source must contain references to source columns in the implementation code or a source datastore be explicitly set in the execution location. Check this mapping code, "

Answer 1

Thank you.
I changed the Execute On option.
Then ODI requests Primary Key for tables; then I add primary keys.
Now I simulated and try to execute; but there is no union opertor in the generated sql.
I checked the address: http://odiexperts.com/tag/union
They are using java code for implementing multiple datasets with different sources.
In some other examples, they are using same tables from different schemes like HR.EMPLOYEES and SCOTT.EMPLOYEES
But I have two tables within the same scheme like hr.employees and hr.managers
Generated sql is like:
insert into EDM_PRD.I$_TARGET_TABLE
NAME,
SURNAME,
ID,
IND_UPDATE
select
HR.EMPLOYEES.NAME,
HR.EMPLOYEES.SURNAME,
HR.EMPLOYEES.ID ,
'I' IND_UPDATE
from HR.MANAGERS MANAGERS
where (1=1)