DSXchange

This has been discussed before, for example here . Essentially, your job is taking too long to start. You need to investigate why this is. What's the elapsed time between the start request and the actual job start (these times should be logged)? There's a hard-wired limit that generates the -14 erro...

The following findings from DS 7.1 on Windows. Finding 1. A non-admin user can not create a UV table. Once that user has been granted CONNECT privilege, that user can create a UV table. Finding 2. A non-admin user can insert rows into a table that was created by a different user, even though no expl...

There's one very big difference from server jobs - you have parallel data streams! Use a Column Generator stage. 1.Open properties for the Column Generator stage 2.Click on the Columns tab 3.Double-click on the positional number of the serial column you are generating. 4.Click on Generator in the Pr...

The command to monitor how many dynamic hashed files are open on the system (that is, the figure that is limited by T30FILE) is ANALYZE.SHM -d You need to be in the UV account to execute ANALYZE.SHM from the TCL prompt, or you can execute analyze.shm -d (or smat -d ) from the operating system comman...

There's nothing changed of which I'm aware. UV tables are, of course, subject to the usual database restrictions; maybe you now need explicitly to grant CONNECT privilege where previously you didn't have to. Don't have access to 7.1 right now, can get access tomorrow (but only on Windows), so will c...

Try checking the "system tables" check box and seeing what happens. There are always system tables! Try using an ODBC trace to record the "conversation" between DataStage and the ODBC driver (and, possibly, DB2 instance). Note, especially, what SQLSetConnectOptions are set, and what happens in the S...

The 1.0 DataStage training was fun! The major exercise was to design a job to particular requirements, the prize was for the shortest running job.

I got disqualified. Most folks' jobs ran 14-20 minutes. Mine finished in five (apparently) because I changed the system clock while it was running.

Give that man a coconut!

(Heck, and we both got - I dare not say "earned" - points for that silly exchange!)

This may simply be caused by the fact that every file on the AS/400 can be viewed as a table - the sheer volume of "table definitions" being returned overwhelms the Import process. Try using filtering either in your DSN setup or in your Import session. It may help.

It was actually a x.1 release, the x.2 coming out three weeks later.

Parallel jobs have custom (Orchestrate), buildop (your own operator, written in C) and wrapper (wrapping a UNIX command) stages. Refer to Chapter 54 of the DataStage Enterprise Edition Parallel Job Developer's Guide for more information. DataStage manuals are in the Docs folder where your DataStage ...

Let me join the advocates of using a Transformer stage. I postulate the following advantages. There may be a future requirement to filter the rows from the text file or otherwise process them. If you believe anything in IT will "never change" you have some hard lessons coming. The Transformer stage ...

Thanks for that feedback. I've not encountered anyone using a NOSUID mount in the past, because I always explain that the DataStage software requires the capability to use setuid in preliminary meetings with those involved. Sometimes, of course, I have to explain why this is (primarily it's because ...

A DataStage job prepares the data and control files for sqlldr. Once that has been completed, the DataStage job has essentially finished all it has to do - you may or may not choose to invoke sqlldr but in no sense other than being a child process is sqlldr under control of DataStage. Therefore, wha...

It seems to me that you've successfully imported the stored procedure definition, the call statement is displayed in the error message. It has successfully been passed through to Oracle for execution; you can tell that this is true because there is a DBMS error code (you can use oerr on the Oracle i...