DSXchange

That sounds like soething on your client PC has changed and a DLL required can either not be found or has been corrupted. What have you done on your PC since this morning? To cross-check, is anyone else working on different client PCs experiencing problems?

Also, you can test the connection outside of DataStage using "msquery.exe" if you have that installed.

After declaring the system DSN and when you try to import the metadata, did you set the filter to "include system tables" and then select the appropriate Excel tab (that will be the name of the sheet tab in Excel followed by a "$")? Do you have column information in that sheet an...

I would write a generic job to do the dataset -> text transfer. The job would declare no columns at all and use RCP, parameterizing the dataset file name and the output file. In the after-job routine I would call the shell command "orchadmin rm {path-and-name of dataset}" to delete the dat...

Are you sure that you haven't switched around your primary and secondary input links? Assume they are switched, would that explain your output behaviour?

No, a user does not need to be an administrator to access the director tool.

For your user, did you fill in the "suite components" in the administrator?

In that case it would seem that you are sorting by Field2 then Field1. Do you have other sample data which can help show the problem?

What data type is field 2? If the data type is varchar() then most collating sequences will sort "*" (ASCII Character 42) after the digits (ASCII Characters 48-57). If the "*" denotes a NULL value in that field, datastage defaults to sorting that before other values. Could you be...

What does your DELETE statement look like. Basically, if the delete is only called one time, i.e. one input row and the delete is "DELETE * from <table> where <condition>;" then it might burst the rollback space regardless of the stage settings and that is what it looks like is happening. ...

The Database has run out of space to (temporarily) store the records in your delete transaction. Either define a smaller transaction size or have your DBA increase the rollback space.

What commit sizes did you experiment with? Which stage are you using on which release?

Have you checked your subsequence log when called from a master sequence to see if the parameter values that you expect are actually being passed in?

Do you have an error handler "ehGenericExceptions"?

My first thought is that the difference might be in the attach point for the process and it might work if you use absolute pathnames. You could check this by adding a "pwd" command to see where your process is attached.

The director program uses the GUI and is not a comand-line oriented program.

It sounds like "dsjob.exe" will do what you want, e.g. "dsjob -ljobs". For a full list of your options with dsjob.exe either call it with "dsjob -help" or see
IBM dsjob description

It looks like your expected output could be achieved with 2 stages: 1. As Craig already mentioned, do a simple horizontal pivot 2. Perform a unique sort on your resultant data on the output Col E If you start with the horizontal pivot stage and get that to function as you expect then you can decide ...

Are you certain that your var1 string contains the single quotes around the values? If so, then just doing ...IN (#var1#)... in the SELECT should work. I don't have DB2 here, but tested it with other connectors and that functioned correctly.

In your post I see double-quote in the error message.