DSXchange

It's the effect of Continue in Lookup stage. Even failed lookups make it to the output, but have columns returned from reference links set to NULL. Just like a left outer join.

Buffers are not static - they are being filled and drained all the time. Therefore the fact that they can contain only 340 records at any one instant should not be a problem. If the consumer can read at the same rate the producer is writing, the resistance mechanism will not come into effect. It's o...

Use a user variable activity to create a user variable at the beginning of the sequence, and pass this variable as a parameter to each of the downstream jobs.

What, precisely, does "it doesn't work" mean?

I'm not sure what you're asking. Ordinarily records are nowhere near 3MB. These buffers exist in pairs (by default) for each link in the job design. The upstream, or producer, stage writes to one buffer while the downstream, or consumer, stage reads from the other. When a certain threshold is reache...

CLEAR.FILE (or CLEAR-FILE) always clears all records. Basically it overwrites the file with /dev/null then replaces the empty header. So can you suggest which records are not being cleared?

Try editor@dsxchange.com (he's usually fairly responsive).

Any time soon - though it IS the weekend.

Any environment variable with APT_ at the beginning of its name applies to the Orchestrate engine (and therefore to parallel jobs only). The same goes for environment variables whose name begins with OSH_. APT comes from "Applied Parallel Technologies", the name of the company that invente...

Make the file name a job parameter. Use that parameter in the Sequential File stage. Populate the "FileName" output column of a Transformer stage using the same job parameter. Of course, this only works for a single file, but you can use a sequence to loop through a set of file names.

I'm guessing you want to set up delimited lists for a StartLoop activity. Americas, Asia Pacific, Europe?

Don't believe that's possible in a Server job.

Probably not on a UNIX-based server.

There isn't one out of the box. You may be able to leverage the Checksum() function, but it's not really the same thing. Or you can write your own. There are plenty of examples out there on the web.

Create a new profile, for example called server1, that can be dedicated to IBM InfoSphere Information Server.