DSXchange

The order of key definition is important to DataStage, sorting on columns A,B,C and then doing a remove duplicates on keys C,B,A will generate a warning message at runtime.

Could you explain what you mean by 'same template'?

In.Col[6,9999]

What is your field's data type? If a string, then "In.Col[1,4]:'0':In.Col[5,2]" would do the trick.

I should note that Klaus, who gave us the information, was working for IBM at the time and could not do anything apart from referring us to the official IBM support path.

What about access rights to the directory and its contents (for the userid that you are running DataStage under)? Try temporarily changing the access to 777 to see if this might be the cause.

Why not declare the field as nullable, then do a dummy NulltoValue() or similar call downstream?

I think that Munch9 has found the cause of your problem. If you are getting 1Gb and divide that by 2097151 that works out to a block size of 512 (a common block size for many UNIX and file system variants). Get your administrator to change that user's maximum file size.

DataStage does not produce a Std out and Std err stream in the classical sense, thus the answer to your query is that you cannot get that information directly. The DataStage equivalent would be the director log file for a job run, which is accessible through API calls.

Change your output sequential file stage format to fixed-width fields and no delimiters.

It looks as if you have put an output row limiter of 1 on that link from your transfrom stage.

Sorry, that was my mistake - I was thinking of Parallel functions and just now realized that you meant a server job. I'm sorry about that. I would write my own routine and use 3 ICONV/OCONV calls: InternalDate = ICONV(InputDate,"D4/YMD[2,2,2]") ;** assuming that is your date string format ...

What do you mean by "System out"?

I'm not sure I follow your description corerctly, but would recommend adding a second output from your transform stage to a peek stage so that you can determine if the erroneous null values happen when reading from SQL server into DataStage or when writing to Oracle.

on a PUT command my guess remains that you don't have enough room in the target directory. What type of UNIX do you have and are both systems running the same OS main version?