DSXchange

Research rsh, it allows you to execute a command on a remote server (Unix or NT). You could have your NT workstation execute an rsh command to a script on the DS server which invokes dsjob.

Search this forum, there's thousands of posts. Here's one:

http://www.dsxchange.com/viewtopic.php?t=85364

Multiple instances are incredibly useful for partitioning datastreams and using multiple copies of the job to process the data. Why have 1 job crunch the numbers on a 16 cpu system when you can have 16 copies of the job with each crunching 1/16th of the input datastream.

Okay, but are you processing ALL target rows every time? Probably not. You probably are processing a vastly insignificant portion of your data each load cycle. Therefore, you only NEED in your hash file the data pertinent to your current load cycle. Consider this approach. Parse your source data fil...

Why is your hash file so big? Are you maintaining a copy of a target table (or subset of columns) in a persistent staging hash file? Why not simply populate the hash file at runtime with those rows necessary for reference? Your hash file is too big for caching, requires significant maintenance inves...

These questions suspiciously look like those that would be asked in a job interview technical assessment....

Routines and functions are stored in a localized file to each project (DS_ROUTINES). In order to make them global to all projects, you will either have to pick a project to house the routines and then destroy the localized file and replace it with a synonym pointing to the global; or, you'll have to...

You can, using very specific steps, setup routines to be globally available to all projects for Server jobs. However, the effort involved IMO is not worth the trouble. Craig has given you the best solution available which has little maintenance effort. To register a function or subroutine to all pro...

I didn't say only one cpu was being used, I said the net effect should be about one cpu. You're only as fast as the slowest piece in a series. CoSort's plugin is available thru them, not Ascential. As for sorting, maybe you could clarify what you're trying to achieve and maybe we could give you some...

By sorter, I assume you mean a Sort stage? If so, then what you see happening is that the sorting is single threaded and happening as fast as the source data can be fed into it, which by a single threaded processes would just about give you a net result of using only a single cpu. Notice 25% of 4 is...

yes

There is no difference in the scheduler, as it is just a GUI over cron. There is no difference in the job logs, they are proprietary and closed. If you need data from the logs, you must extract it using the published APIs.

Use DS BASIC job control in an Batch job. Setup a loop to cycle looking in the directory of interest and execute the jobstream for each file present. As you finish a file, either move it or rename it to take it out of the processing loop. Continuously loop until the list of files are exhausted. You ...

Read the data from the file as a single column. Use stage variables to queue data and output on change of group --> the "}" character. Since your last row should be that } character you shouldn't have the issue of the last group not being output.

DataStage has poor encryption capabilities. For example, even encrypting a password doesn't mean you can't expose the unencrypted value. Just simply map that parameter into an output file as a column derivation and you will see the value. One useful method for passing in the parameters is to have jo...