DSXchange

Additional stages will consume additional resource.

If your concern is about input, its better to have two streme, one for first row with lookup and another for rest of the rows for clarity purpose. You can also build 'If then else' logic with INROWNUM=1 in the derivation part. Or you can had constraint inthe lookup.

While inserting into the table, the value for the Sequence number column would be seq.nextval, where 'seq' is the sequence created using CREATE SEQUENCE statement.

If the Descriptor file has been deleted using rm command, then you have no choice left out, other than deleting the data file using rm command. But as mentiond, you need to be carfull about the data file that do not belong to any other descripot file. Optionally, you can try to create another dummy ...

What are you expecting? What is not happening? what is the error you getting?
You will see the Double quotes only at the output/input file. If the quote character is given properly, you dont need to worry about it.

Untill you have the same set of columns been propagated from the source till target, RCP shouldnt make any difference.

As suggested much early to check for disk space. Nice that you got your problem solved.:D

If you are editing the exisitng column, you need to edit the existing user defiled code as well.

I guess, they are messing up with "Informatica" stages, and referring to the sparse lookup.
If the number of rows in the two sorce is not extremely huge, explain them that, lookup made through the hashed file is more effecient that an sparce lookup done over an indexed table.

Read the parameter in the JobSequence that useses the CommandActivity by assigining the default value as $PROJDEF. And now you can pass the parameter to the commandActivity stage from the JobSequence.

It seems, sequential file cannot hold that huge numbe of records in the buffer.
Try to check the your project directory disk space, if it can hold the file with the same size of the input file.

Protected project cannot be compiled by users other than dsadm. As Ray said, if a user is given access to compile, the user as full access to edit the code.

If the number of records for each group will be only 3, then its possible by using mod function and some corelated subquery. But it is more simple to accomplish from datastage.

What is the output datatype. If it is char/varchar, juct Convert '-' to ''

Convert all the statement with in the cursor in the SQL into a single record set in Datastage, may be by using single or multiple jobs. User Transformer, for all you typecasing, case condition, If then else check... If it is PX job, leverage the filter, SWITCH for these purpose. Now cursor can be tr...