DSXchange

mgsteiner wrote:Please can you make visible the entire message to me, because I'm still not a Premium member.

That kind of defeats the whole purpose of setting up a system like that.

Well... the variable length record part with packed decimals and a record length field at the front would work in the CFF stage - up until the 'padded to a multiple of 4' part, which throws a damper on things. I'm thinking you'd have to go 'BuildOp' (or some other custom solution) here but curious w...

So... what's your workaround?

Many of the places where I've worked preferred an 'all or nothing' like you are describing. We managed this back in the Server world by simply setting the Transaction Size to zero which meant a single commit is issued at the end of the load. Yes, that does mean you need to work with your DBA team to...

DSUser2000 wrote:(no idea if that is any standard what you are describing)

EBCDIC Zoned Decimal

Either import the file into your database and join them there or export the table to a file and use the Merge stage. From what I recall,those are your two primary choices in a Server job. Depending on the type of kind of join, you may be able to put the file into a hashed file and use a lookup to pe...

That would be the Merge stage, not to be confused with the Parallel stage off the same name.

I don't but I'm sure you can google something up rather easily. I'm not sure off the top of my head that you can do all of this from 'your end'... I'd imagine your company / SysAdmin would need to set this up on their end or at least be involved in the process. [shrug]

Everything runs on and all work is local to the ETL server, so things like files must be there to be read and will be written there. The 'workaround' is the use of an NFS mount (or Samba and the like) so that external systems appear local to the ETL server.

Odd... but good to know. Thanks.

What kind of conversion are you doing in the transformer - implicit or explicit?

Nope.

Sure, just use the proper field delimiter - a comma in that case.

A UNIX script. And if you really need a DataStage job to run it, use a Sequence job. However, what does "local machine" mean here - the ETL server?

Start by posting in the correct forum. :wink: I moved your post. Secondly, provide more details as I don't really see any of those stages requiring any 'performance tuning'. Did you have a specific job or job design in mind when posted this question? And since you started in a forum that doesn't req...