DSXchange

Curiously I answered this exact question yesterday.
viewtopic.php?t=126771

Your job design includes some kind of delete - you need to find that, and check the file itself in the operating system to find out why it can't be unlinked. Maybe it's in use by another process/job? DSD.RUN is the program that runs a server job or a job sequence. Since you're in the parallel forum ...

Actually, the proof is in the score.

The orchadmin utility can delete a File Set - maybe it can dump/copy File Sets as well? Just provide a control file name with a ".fs" suffix. Note that this is a guess, try it and let us know.

The big one is the reference input to a Lookup stage when in a multiple machine environment. Entire (which Auto gives in this case) has to send all reference rows to all partitions. Better is to partition both stream and reference inputs using a key-based algorithm using the join keys.

Only if you can guarantee that the job name will appear only in the process(es) involved in running the job. What if, for example, the job design were being monitored or open for editing in Designer? Even so, I regard that approach as a kludge - prefer to create a loop and check the individual jobs/...

DataStage/SQL does not support a CASE structure.

You can, as Kim noted, use an I-descriptor or EVAL field modifier to build a nested If..Then..Else expression that would be equivalent.

Depending on your need, and whether there are any columns dependent on just one column of the primary key, you may be able to get away with just the one table, with one row for each order. You can select distinct vendor from this as required.

Use a format mask.
Reject records with the rightmost two characters >= "60". A constraint expression on the "good" output could be:

There was a recent webinar on precisely this topic.
Go to www.dsxchange.net/webinar_training.html and look for the title "Files and DataStage Job Sequences"

Show us all the properties of your StartLoop activity (there are three) and the design of your job sequence. Without those your last question is impossible to answer (apart from "you're probably doing it wrong").

The do have it, it was shown at IOD 2008. Whether it's "available" is a different question. If your support provider can't find any information (after being leaned upon in a suitable fashion) please post back here, so some of the IBMers who are involved can be alerted (and maybe even post ...

Start by dumping and examining the score, so you can figure out which operator player 6 is.

Search the forum for the exact message about the debugging utilities gdb and sed.

karry450 wrote:I am writing flat file to oracle using Bul Loader

Ah, yes, so many projects end up using a Bul Loader.

Use a database that can handle multi-valued fields, such as UniVerse, Unidata, jBase, D3, etc.