DSXchange

Of course you can recreate the directory, just make sure you match the ownership and permissions if you do. And don't "clear" it if any jobs are running unless you stick with the "x days old" approach that Ray already posted.

Do not run the jobs in the background or use the O/S wait. Do add the "-wait" option to the dsjob command.

Why not use a Sequence job with the Start/End Loop stages?

Silly solution? More like an absolute requirement.

I thought (from other posts here) that DB2 wanted dates in internal format.

From where? What does "calling SQL stmt" mean? Shell scripts run from the command line, not a TCL prompt. Are you attempting to use your database's CLI, such as sqlplus for Oracle?

You haven't given us much to work with here.

What is your target database?

So... in other words... one didn't show up.

The & is a shell metacharacter so you need to escape them as Ray did.

For whatever it is worth, that error code means:

0x40EC0, "Fetching requested number of rows will exceed total number of active rows supported by the rowset."

What exactly are you using to "see" the output?

evee1 wrote:If I change the contents of tnsnames.ora do I need to restart the Designer, or does it reads the ora file on request?

You don't need to restart anything after that change.

Then either change "-maxdepth 1" to "-depth" or remove it all together as it really isn't needed for this task.

That option may be platform specific, check your man page as I suggested. And Ray missed the "where to search" part, add a "dot" right after the find and see if that helps: find . \&PH\& -maxdepth 1 -type f -mtime +10 -exec rm {} \; Assuming you are parked where it should...

From the "command line" of the DataStage server, i.e. from your operating system. Both istool and dsjob can be run from there.

From the command line: "man find"... and basically "yes" to the second question.