DSXchange

You didn't ask a question. If you search the forum you will find that this message is perfectly proper behaviour of DataStage.

Are you trying to view unconverted EBCDIC value?

Welcome aboard. :D

The Merge stage in server jobs performs a standard join between two sequential files based on a common key. Sounds ideal for what you require.

Makes more efficient use of memory if the two files can be pre-sorted.

A standard, local (not in COMMON), scalar variable in each case.

"?" is NOT the invalid character. "?" is a special Unicode character that DataStages uses when it can not convert the invalid character into Unicode (its own UTF-8 encoding of Unicode) based on your assertion of how the external character set is encoded. You must determine precisely under what stand...

Most likely more physical memory, or more virtual memory, or higher ulimit setting for the executing user.

Without DSDetachJob, the fact that the job handle variable is in COMMON will cause an attempt to "remember" the job to which it was attached. I would recommend not using COMMON variables for job handle variables, and always to detach any attached job before exiting from a routine that attaches it, s...

Different sets of characters are encoded according to different standards. For example, Chinese characters may be encoded according to standards called BIG5 or GB2312. There are many different standards for encoding Japanese characters. DataStage provides maps for most of the commonly-encountered st...

"?" represents a character that could not be mapped.

The correct way to get rid of it is to choose a map that correctly maps all the characters in your data.

Do you DSDetachJob near the end of the routine? The fact that you have the job handles as COMMON variables may cause a failed attach on the next run. As a general rule, never have white space between a function name and its left parenthesis. It sometimes works but is never guaranteed to do so. You d...

How large is the hashed file? Have you checked that it is structurally intact?

You can run them anywhere. Post your routine code and we'll take a look.

For example if you'd used DSJ.ME as the job handle, that would have returned -1 if run in a Routine activity. But you've already said you attach the job.

The ulimit setting for the user ID under which DataStage jobs run must be set to unlimited by the UNIX Sys Admin.

Why not build both and publish your results?

How good are you at writing efficient DataStage BASIC routines?

Welcome aboard. :D

I have found that it's sometimes necessary to re-boot the Windows server after changing an environment variable for that change to take effect for background processes (such as those that run DataStage jobs).