DSXchange

Is this a server job (as your header) or a PX job (as in the forum you posted it in)? What does your shared container do? Does it contain sequential file stages which write to a file that might be the same for concurrent runs, i.e. if the name is hardcoded or as a shared container parameter that oth...

Which stage are you using and could you perhaps cut-and-paste the error message? I don't think that any manufacturer would actually put "due to a defect in our product this won't work" type of message in their product. If you do have an array size then the commit frequency does need to be a multiple...

I can't help you there - just that it is a known issue that IBM/Ascential support is working on.

Ramesh, DataStage has no idea what hardware lies at the base of the system, the lowest virtual machine layer that it sees is the OS. In this case the OS remains the same with AIX and thus going from one p-series to another is transparent to DataStage. If you mean LPARs and dynamic switching of CPUs ...

The job type resides in a column in the DS_JOBS file. I don't know of any published API routine such as DSGetJobInfo() whre this information is available, so the only way to get this would be to read the DS_JOBS file - and that will only work until the repository is changed at version 8. You can eit...

As I stated before, if you have a partitioned DB/2 database and are using the LOAD method then your problems are most likely caused by a bug and you should contact your support provider.

That command will solve the load pending/backup pending. I think it needs to be issue for the session that is doing the bulk load. If you don't use the bulk load but the user SQL then the job will go through correctly (perhaps not as fast as you can get with a bulk load, but it is better than what y...

Hyd, Google is your best friend for this. Sorting data is easy. Any introduction programming class starts off with a simple "bubble sort" and most of the time things stop there, since doing an efficient sort is a very mathematically complex thing. The built-in sort for UNIX is a very efficient algor...

In most cases the default partitioning method is sufficient and the best to work with. If you are writing to (or reading from) a partitioned database then that is the most efficient partitioning scheme to use in the job as there is no necessity to repartition. This topic is covered in the documentat...

I'm a bit confused now. If your database is partitioned and you are using "LOAD" then there is an issue with DataStage (we discovered that last week and have an eCase open). You can avoid your problem with the backup pending state by issuing a db2 set integrity for DB2Schema.DB2TableName immediate c...

What is your write method in the stage (if it is TABLE, try generated SQL). Also, have you tried running on just a single node (i.e. in sequential mode). Turn on your APT_DUMP_SCORE to see if that tells you anything more. Also, is the DB/2 database partitioned?

A script is usually not necessary, just "dsjob -run -mode NORMAL -wait {project} {job}". I f you use the builtin SEARCH facility in this forum you will find many posted scripts that will give you good working templates. Some are simple, others are complex and do all sorts of error handling.

Can you cast this as varchar(32) and define it as varchar(32) in the datastage column definition? I don't have access to Informix here. I am still surprised at your error message, since that is generated in SQL and not in datastage it doesn't make sense in this context.

dwblore wrote:..WHOLE message quoted

There really is no reason to quote an entire post, particularly if it is the one right before yours. It keeps the pages to a manageable size and makes it easier to read and understand.

Named pipes are usually used for simple interprocess communication. DataStage supports named pipes.