DSXchange

As you've already noted, using buffering can result in increased throughput. If a machine is memory-constrained then using additional buffering would exasperate the situation instead of speeding up processing, thus one is given the choice of turning it on or off.

Generally I would say that "no", you shouldn't have problems with server resources.

The join stage will first sort the input streams according to your join key(s), unless that has been already done or explicitly turned off. That way the join mechanism can process the data without having to store large amounts of interim data in memory or on disk. Each Node is completely distinct an...

Put in a user variable stage, create a variable there derived from your counter (i.e. RIGHT('000':$counter,3) ) then use that variable as the parameter to your called job.

What directory is "sqlcmd" in? Is it in the runtime $PATH? Note that the PATH is changed in the "dsenv" file and this might be where your error is being introduced.

The answer to the backup frequency question doesn't depend upon the technology, but on the answer to the question "How much data loss can we afford?" and, concerning incremental backups, "How much downtime can we afford?" as incremental restores tend to take longer than full rest...

Unofrtunately there is no builtin timezone handling is DataStage PX, once would need to code that manually. This would be easy if it weren't for daylight savings time.

p.s. You can delete your duplicate post as long as nobody has replied to it yet.

Chances are very, very high that your job's bottleneck is not in reading a hashed file but somewhere else, thus changing buffer and cache sizes, if possible, would not make a difference. Are you asking about reference reads, or perhaps writes? Also, perhaps you might be looking at public/private lin...

Same problem as before, the path to your "sqlcmd" program is not in your $PATH variable. Either fix the $PATH or put an absolute path to your program in your bat file.

DataStage will do an implicit conversion, but it is much better to ensure that the metadata is identical and do an explicit conversion. Char() field conversion are always blank padded. I don't know the rule for conversion, perhaps DS uses the left link as the original and converts the other links to...

If the HTML page has a valid XML structure, then the XML input stage might work for you. Could you post a sample of the page contents?

That message happens when a "RESIZE" is done on a hashed file. Another possibility is that you have run out of T30FILE units, but usually that is accompanied by other error messages. Does the error always happen or is it sporadic? Are you calling any routines in your transform? If you post...

Can you specify one sequential stage per file then run them all into a funnel stage where you specify that the data is already sorted?

DataStage is designed so that when a job aborts it cannot be re-run again until it is reset. This is intentional so that a job cannot just continue to abort on subsequent runs without any noticing that it doesn't work. The "reset" of a job does nnot perform more than changing a status flag...

typing in the Euro explicitly should be OK. What method are you using in your Oracle stage - insert/upsert or table load, that might be where the problem is creeping in.