DSXchange

They're all viable. Prefer the DB2 Enterprise for parallel write to partitioned tables. Otherwise experiment and determine what works best on your particular hardware and software combination. As a general rule direct write or bulk load will load data faster than SQL-based approaches.

Use ipcs -m | grep adeb to identify the users. The shared memory key's final four digits are the "UniVerse" user ID in hex. Tip: use ipcs -m | head -1 ; ipcs -m | grep adeb to include column headings. Use ${DSHOME}/bin/list_readu | grep ADMIN to identify which projects are in use. Each con...

Define "performance".

It depends what you are trying to do. Each stage type has its own particular task(s) to perform. For example, the DB2 Enterprise stage can not (in version 7.5) access DB2 on AS/400 or mainframe platforms. You use the DB2 API stage in this case.

If you already have a timestamp, which I infer from your original post in this thread, simply convert it to a string, then use substring and concatenation to build the new format. Converting all the internal delimiters to "" first will make this easier. Assign stage variable svTS with Convert(&q...

If you already have a timestamp, which I infer from your original post in this thread, simply convert it to a string, then use substring and concatenation to build the new format. Converting all the internal delimiters to "" first will make this easier. Assign stage variable svTS with Convert(&q...

If your client is running Windows XP sp 2 do you have the DataStage xpsp2 patch installed?

Can you be more specific about your use of the term "parameter"? Is this a job parameter, or do you simply need the current date within your job design? Either is possible, but we'd prefer to answer the correct question.

Indeed, such tools are jealouly guarded by IBM, and only available as services (that is, you never get to see the tools and they're never deployed to your hardware). As to whether they've developed one for Ab Initio, I have no idea. Why not ask IBM, specifically the Center of Excellence for Data Int...

Welcome aboard. In parallel jobs the Merge stage does not read from source (unlike its server job counterpart). Upstream of the Merge stage you will need appropriate database stages (for example DB2 Enterprise stages) to read the data. You also need to partition and sort the data on the "join key(s)...

You are running out of both memory and communication resources. This suggests that your server is overloaded.

This is a Transformer stage. I suggest you've created and compiled it elsewhere, then deployed it to the current project. Try force compiling the job in the current project.

If all that happens is a hang you're lucky - in some circumstances running out of disk space (for example in resources disk or scratch disk) can be fatal. Solaris is notorious for configuring a tiny /tmp - you should try to use a different area for temporary files (see TEMP environment variable and ...

Your job does not have a job parameter called "file". This is what is generating the error when you use this syntax, which is referring to a job parameter called "file".

Prove (perhaps using echo commands) that all the environment variables you have specified on the command line actually have values assigned.

What are you giving as the invocation ID values?