DSXchange

You need to separate the column names with a comma in the KEEP statement in a modify stage. Your line needs to read "KEEP ReceivingDocumentID, CreationTimestamp"

Since Linux is just another flavor of UNIX there is little difference in performance based on the OS. The major differences will be noticed with the underlying hardware. There should be no problem 'migrating' from one to the other-.

Go to the director log for the job. Right click on a line with the warning message and then choose the option "add rule to message handler". In that window, choose to demote instead of suppress the message and click on "add rule".

There is no way to set or change the auto-purge settings through the command line. The only documented way is through the graphical interface in the director on a per-job basis or by setting the defaults in the Administrator tool. There are undocumented methods of changing the autopurge settings and...

You threw me off with FMT reference, since that isn't applicable to "view data". You cannot right justify in that window, the only formatting DS really uses is the field width to presize the display width.

Do you mean with "view data" in the designer or from somewhere else? The FMT function will work from the command line, but has nothing to do with "view data", where the datatypes and widths play a role in how things are displayed.

When you convert your .xls into a .csv make sure that you quote all strings, then you won't get this problem.

The fastest way to kill a job is to turn off the power to the server The only officially sanctioned ways to stop a job are: a) use the "stop" button in the director or from a DS/Basic call to a job you have started. b) if (a) fails after a given period of time use the "cleanup resources" menu tab in...

If you actually cut-and-paste the error message into a search window here on DSXchange you will find many answers.

Start with a simple case. Declare the file as pipe delimited with no quotes and set up the correct number of VarChar columns. You should be able to "view data" without errors or warnings. Once you have that working, you can start declaring the columns with the correct metadata. When you post problem...

Nothing like that is cached in DataStage. Have you had project corruption issues in the past? If not, then I think that copying the job to a new name and seeing it run correctly was probably coincidence. Does the problem still exist with the old program or does that work correctly now?

The code I gave earlier goes into the Sequence job and the new date-time is passed as a parameter, then you use the parameter value in your source stage.

The parameter value cannot be computed in the call itself, you need to do the math before passing the value. If you are calling your job from a sequence you can do this in user variable derivation as below: WorkTime = ICONV(Field(OrigTimeStamp,' ',2),'MTS')-(60*2.5) NewDate =...

I cannot recall exactly what we ended up doing (after all, it has been almost 2 years), but I do remember that I was able to access the data when a view was applied with a char instead of varchar columns.

G SHIVARANJANI wrote:...When i tried to pause the job...

DataStage has no facility for pausing jobs. I don't know what you did, but it caused the job to abort.