DSXchange

What is the database datatype, not the one in the job? Anything CHAR datatype always has space padding, so it's hard to believe that you could pull back more characters than the database allows. Could the job be out of step with the database?

I like using basename, but since it's a Windoze platform that option is out.

That's exactly what I'm talking about. Copy the files from the mapped directory to a local hard drive and process the data there. Then, copy it to a mapped directory.

You need to declare a COMMON variable, which will maintain its value as long as the process is active.

COMMON /pickaname/ yourvariable1, yourvariable2, yourvariable3.

Search the forum for implications and limitations to a labeled COMMON.

You'll need to separate your inserts from your updates during transformation process so that you have a link for each. Then, you can get job link statistics using DS BASIC API calls or the dsjob command line program.

Go to my site, request access. I'm here for another 15 minutes so I'll approve it right away.

You don't have to change jobs. Just copy files around. If you know how to map network drive on a Windoze PC, that's all you need to do on the server.

Count the slashes and the filename is after the last slash. SlashCount=COUNT(yourstring,"\") If SlashCount > 0 Then LastWord=FIELD(yourstring, "\", SlashCount+1) End Else LastWord=yourstring End If RIGHT(LastWord,1)='"' Then LastWord=LastWord[1,LEN...

The DataStage server has to have visibility to all network drives in order for the jobs to access them. When a job runs, it runs as the user logged into Director but still on that Server. Any network drives should be mapped on the server, else use UNC notation when accessing. You want to avoid doing...

If you grab my mass-compile utility off the members area, that includes a CLEAR.FILE of the logs when it recompiles the jobs. I didn't write it to preserve the purge settings, probably should fix it. Maybe you could make a better version? There's no 1 line command to clear a bunch of hashed files. I...

Correct? Yes, for a minimum out of effort. I'm in the camp of spooling out the entire row (all columns) and staging into a sequential file first. This allows you to have a "before" image of a row that may be potentially updated. Then, another job loads that sequential file into two hashed files: on...

A sourcing, extracting, unloading, exporting, or spooling job would be something like OCI-->XFM-->SEQ.

Spooling, old slang for extracting the rows from a data source and writing to a target. "Spool to tape, spool to disk, spool to file." In this case, I'm talking about writing out to either a sequential or hashed file from a database. One would never spool to a database, however.

CLEAR.FILE is not the same as DELETE.FILE.

You could get your hands on the utility I've got that mass-sets the purge settings on jobs. You could use the other utility I've got to go mass-purge all job logs, then re-apply the purge settings.

Okay, I just freehanded the junk, F and V are close together on the keyboard.