DSXchange

The scheduler used on the Windoze platform is the native Scheduler service on the server control panel. You'll need to look at any third party tools or Google to see if anyone has done something like you're asking. I will tell you, the Windoze Scheduler is not an Enterprise Scheduler, so it has none...

Yeah, one-time... How many "one-time" processes go live and form the code base? Do it right the first time, blah blah blah. Besides, it's posted here, what's there to do other than tweak to taste?

Sure, use the AFTER-TRANSFORMER routine calls as well. Just pick the transformers that execute at the end of the job logic branches. Sometimes you have transformers closing at different times, because of how you constructed your job.

Overkill??? I like to think full-featured

Well, a single instance of a job reading a seq file, doing a couple of ODBC lookups, and writing to a seq file will not use multiple CPUs to run. By instantiating the job and using a constraint to partition the source data, you can run as many instances as you have CPUs and increase your net through...

Anyone use the conversion script for turning 8i plugins into 9i? I'm about to help a customer and we were going to use it to hammer out about 500 jobs. Any glitches or gotchas?

That's just the way things work. You can try displaying fewer columns and then streeeetching the 400 character column really wide.

It's stupid because it's trying to be smart and fails miserably. The toolmakers insist on foisting internal date requirements eradically/sporadically throughout the tool. Some stages need it, others don't. Imagine you have an Oracle OCI stage writing to an aggregator then to a sequential file. In or...

Because the aggregator is stupid. Set your metadata to string or "internalize" your date using the ICONV function, then externalize the date after the aggregator using OCONV. The easiest is to just set the metadata to string if your date is guaranteed in ISO format.

FYI, the DS BASIC APIs that retrieve log file rows from the main or individual instance job have the same issue. It returns all log rows for all instances, you'll have to loop thru all rows and parse for your jobname.invocationid rows.

You're better off using TIMEDATE(), as it give you a synchronized date and time value, preventing any midnight math issues. If you use DATE() and TIME() with separate calls, you could accidentally span midnight at that blue moon moment right between. Just parse out the date and time information and ...

Your questions are part of a philosophical discussion common in ETL circles. There is a group of followers, including myself, who advocate "off-target" transformation. I believe in transforming your data within a "sandbox" area to complete resolution, especially when there are multiple repeating row...

You have to write a routine that "does something". In that DS BASIC routine, do whatever you need to do: open files, look at link counts, log a message, send an email, whatever. Just getting the link count does nothing, what do you want to do about it? So, go into DS Manager, write a routine, and us...

Have you seen this post?

http://www.dsxchange.com/viewtopic.php? ... dsmigrator

If the field is blank it will pad it with whatever you want. I don't know how you got the error mesasge, but you can supply the 3-digit ASCII number for fields when specifying delimiters or pad characters or quote characters etc. 032 is a space, 009 is a tab, blah blah blah