DSXchange

No, not legally.

So your rule is "if the date and the ro_code are the same, then do a logical "OR" on the column values"? Is that correct?

I don't know about the SCD stage, but when I use GetNextSurrogateKey() in a transform stage at V8 on AIX I do not get an error.

N.B. have you tried compiling even when the link is red? It might just reflect an issue with your project DSParams file.

If you add a copy or transform stage does the error go away? Can you do a "view data" on the dataset? Can you tell me where you got the SQL Server stage, I haven't seen that stage before and would like to use it as well.

What exact derivations have you used, please copy & paste them since the data in your post isn't complete.

The rows aren't identical, some of the columns are. You haven't described what the combined row is to look like or what the combination rules are.
This type of thing is commonly done in DataStage.

I'm not at DS to confirm that. Why not bypass this problem and use a derivation of "IF IsNull(In.Column) THEN 'Value' ELSE In.Column"?

Sud - it is greyed out in the administrator in many places if you are not the admin user even when NLS is installed.

Look for NullToValue in the null handling tab of functions, you will find it there.

Did you fill out the details in the Surrogate Key tab in the transform stage?

There will be an NLS line in the "smat -t" command. If the value is 0 then no NLS, if the value is 1 then NLS is installed.

There are other ways to find out if NLS is installed, if you develop a job and the NLS options are greyed out then it most likely isn't installed. I'm not at a DS system now to find a specific example but chances are very high that you don't have NLS installed. Does your dsenv file set NLS_PATH? If ...

The commented-out lines are OK and are even commented out in correct NLS installations. If you login to the admin program as DS administrator and the NLS is greyed out then you don't have NLS installed and the NLS parameter are not used.

Jobs are started from the director or command line, so it is a simple matter to start several at the same time; or make one multi-instance and start several parallel runs of the same job. The difficulty lies in creating the job(s) so that no locking or other problems occurs and from you post it isn'...

I would write a sequence, call the SQL via the command line, parse the result into variables in the sequence and then pass those values to the jobs as parameters.

EDIT : I missed that you needed to do this for every source line. In that case, yes you can used a stored procedure for this.