DSXchange

Thank you much, you are right I tested with only key 1 and key 2 and it worked just fine, removing the duplicates. I guess I has to use remove duplicate stage and retain the first record. After the transformer stage I will use the same partitioning in the remove duplicate stage and retain the first...

I have a requirement in which if a duplicate value occurs in a particular field, I need to drop both (or all) of the records with that value. I'm guessing there may be a sql statement that could be devised to handle this, but I'd like to do it without that if possible. Anyone have suggestions? Than...

srinivas.g wrote:i ran same thing in datastage 7.5.1 A with four nodes.

But it is exceutes only once.

the close command query is

BEGIN
DELETE from <table name>;
COMMIT;
END;

try to insert into the table and have a primary key.
job will abort.

Regards.

[quote="DSguru2B"]That is wierd. Can someone reproduce this on 7.x ? Get in touch with IBM and see what they have to say about it.[/quote] it happened to me as well. i was inserting into some table on 4 nodes, and my job aborts saying " i was trying to insert duplicates" . i am o...

thought diffrent column names on each query...you were right

check the column names you gave in the colums tab. you are using union.. in the first query the first column is emp.a in the second query it is emp1.a use a common column name like this select emp.a as a ,emp.b as b ,emp.c as c from table emp union select emp1.a as a ,emp1.b as b ,emp1.c as c from t...

I this is happening since u have Filed Marks in the Execute Command output...

when assigning the output to the Job parameter, remove the the FM.

were you able to run your query on DB without any issues???

are you writing file to file??? or from file to a DB table???

does the SP runs with no issues when you run on the DB directly???

you can write that data set to a seqential file on prod and copy it to test.

before you run the job write from seq file to a dataset on test.

may be an option atleast.

yes you can go for that...unless you want to mention INSERT/UPDATE statements manually.

then write to a dataset load the DB eventually.

[quote="chulett"]It other words, do a lookup against the target table and only load those where the lookup fails. ...[/quote]

EXACTLY

you can use WRITE METHOD TO "load" if you just want to insert. but if you want to reject the records those exists, you can't have reject records tapping option in the ORACLE stage( in LOAD mode). you may have to do it in your job seperately and reject the existing records before you load t...