DSXchange

Oconv(InLink.TheString, "MCP") will convert all non-printable characters to ".".

Yes. Employ Harry Potter.

Welcome aboard. I can only suggest that your POC is ill-designed. Grab three hashed files from the target tables. One contains a single record, with a dummy key and the highest current key value. The second is a reverse lookup from the business key (however many columns there are in it) to the surro...

No idea at all. How about you tell us precisely HOW you are attempting to invoke sqllder from DSExecute, and what error messages you get?

Neither is Convert(). I thought you may be considering other stage types. Can you post how you are attempting to merge the rows with the Row Merger stage?

I understand that you must use the XLC compiler on AIX.

Use a stage variable to accumulate via concatenation and constrain the transformer stage to emit output only every third row.

If you are systematic about it, it's not that great an obstacle.
I've been doing exactly that for 12 years with DataStage.
And it means that, where the transformation logic differs even slightly, you don't have a problem.
Or if the rules change later, which they have a nasty habit of doing.

Maybe you could more simply have multiplied by 100.

Can you give an example with values? If even only one row is returned from the lookup reference link, then the lookup is considered to have succeeded.

Yes of course. By using an Aggregator stage. Or by using pre-sorted data.

One can not comment about whether or not some technique will affect performace within an ETL context until one understands - has a definition of - what performace actually means in the ETL context. So, if you would care to document your definition, then perhaps we might be able to offer some cogent ...