DSXchange

Not quite sure all what would be considered 'invalid' by the stage but the docs say this: This is an optional property and only appears for the Load Write Method. Use it to specify any command, in single quotes, to be parsed and executed by the Oracle database on all processing nodes after the stage...

Well... didn't even need to look to see who posted something with that subject.

Hmmm... what NLS_LANG setting (i.e. characterset) is your job using? UTF8? Want to ensure it supports those characters.

Finally... we get the full picture of both the error and the job design that is causing it. Posting the complete job design in your initial statement of the problem could have gotten us here quite a bit sooner. :? The index may help, but since your picture shows two separate streams embedded in one ...

In addition to what Ray posted - what database is this?

Gotcha - read first, ask parsing questions later.

Not very clear on your question or requirement. Are you saying that you need to run your posted sql 40,000 times - once for each combination of ids on file? Can you not just join the two?

Simpler to build three Sequence jobs that just do the one thing and schedule each at the appropriate time but what you posted is perfectly feasible. One way - build a routine to pull the current hour from the system time and use it in a Routine Activity to start off the Sequence. Branch to one of th...

What kind of "server crash" are we talking about here? We would routinely run quite a number of jobs simultaneously, but of course that is dependant on hardware config, job design, etc etc - and mine have always been UNIX platforms so not so sure of the capacity of a Windows-based solution...

Gonna need some metadata for that.

Dat file? Digital Audio Tape? Something with a .dat extension?

Have you tried to import the metadata from it or have you been given the same? Can you post some sample records?

OK, that's one example. So... the next time you see a file from SAS, could the data / column names / column order / number of columns be completely different? If so, are you thinking you can somehow build one job that could read these changing files automagically? Or is there a finite combination of...

Change how? Can you provide specific examples? Without specifics, people won't really be able to help all that much... except for the 'skipping' part.

So... you need to skip the first, second and last records in the file? Or is there more to it than that... are the column names or number of columns dynamic perhaps?

I would wager it is complaining about the total length of your parameter values, rather than the number of them. So... how long are the values? You can certainly pass more than 5 parameters, btw.