DSXchange

It can also be affected by your CREATE TABLE statement. For example you can define a column explicitly as VARCHAR2(30 BYTES) - which would store 10 multibyte characters - or as VARCHAR2(10 CHAR). There is a global setting to provide the default meaning but I can't recall at the moment what it is.

I use dscmdexport (successfully) with the /D=domain switch with no reference to the port number at all.

Specify format in schema file.

There are many examples in the manuals; format is specified the same way as for non-partial definitions. Inspect one of your extant sequential file definitions in the Repository, particularly the parallel layout thereof.

Check the ulimit settings for the user ID that executes DataStage jobs.

What is "fd_set"?

Please post the entire error message, and identify any stage names (with stage type) and/or link names in it.

It is not an issue of enablement, it is an issue of licensing. Enterprise Edition gets you parallel job and server job licences, but it is possible only to licence parallel jobs. (You're still paying for server job licence, but under no obligation to use it. Personally I think that's a dumb choice b...

Unless you are doing a sparse lookup, the data are always loaded into a (virtual) Data Set. You don't have to do anything. This just happens. Loading into a (physical) Data Set won't change performance at all in the big picture, but may time-shift the cost of loading the data set to an earlier phase...

Yes, put it in the user-defined area. There are many environment variables not in the project "out of the box" - for most of them you discover the name when reading the manuals.

Double "P" in "suppress".

Was the install with NLS enabled?

The first date is YMD order. Is the second date also YMD order? If not, what is the order of the day, month and year components, and have you informed Oracle that such is the case?

US7ASCII is not multibyte. You may be able to read this data with a map of NONE.

Users are typically created by unskilled labor with the job title "parents".

Do you mean user IDs?

Your job failed when attempting to write to a log.

What error message was generated?

How large is the log for the job?

Since your Oracle is set up for Japanese, it uses three bytes per character for storage. This the source of the trebling. It is not a bug.