DSXchange

No, that's simply how it works.
The Lookup stage by default returns one match. If you want more than one (which is likely less of the time) you have to tell it so.

One sequence can call another to build a hierarchy of control.

Probably not in the sense that you mean, though it IS possible to implement "dynamic arrays" which are delimited strings with reserved delimiter characters that pretend to be arrays.

... or using something like QualityStage to figure it out based on known first names. But even that's not 100%. Can't be, because parents go on inventing new names for their children.

You probably have a heap of orphaned *.ds and/or *.fs files hanging around now.

How, precisely, do you go about "Creating dataset in a user defined path"? If all you create there is the descriptor file, you're not accomplishing much. The data, or "segment", files that comprise the data set are stored in the directories identified as "resource disk"...

If file has one header and one trailer this is a trivial task to script. If the file might have multiple headers and trailers, the best solution is probably a DataStage job.

Subtract date-to-Julian-day conversions.

So long as you're not interested in performance. What are you looking up against?

You can install Fix Pack 1 on any supported platform. Yours is a supported platform - the fact that you have installed 8.5 there is not sufficient evidence of this - you need to check the Planning, Installation and Configuration Guide.

You also need to incorporate a mechanism for shutting down the "eternal" job for those occasions when you need to stop the Engine.

<rant>Resist stupid requirements!</rant>
Please post your client's reasons for "needing" it.
Does your client have pointy hair?