DSXchange

Log into manager, click on File, click on New Server Routine. You have an option for Transform Routine or before/after. Choose the appropriate one and start coding.

You know what. Craig is right. VSAM cannot be accessed by sequential file stage. The files that I have at my current site are flat files with binary data (which triggered my suggestion) and not VSAM files. So, forget my suggestion altogether. Wait for someone to come by and shed some light on this. ...

My bad. Not the String, but under General, provide the Character Set as EBCDIC and Data Format as binary.
Along with that, set the Record Length as fixed and Final Delimiter as end.
Set up the each column properly in the column tab (I mean the metadata).

What options did you try in px sequential file stage? Did you try, Format tab, under 'Defaults', String ----> Export EBCIDIC as ASCII ?

Delete the job. Create a new one. Use odbc stage, just for testing purposes. See if the error persists. This is really odd. You said you have other jobs that connect to the same database without any issues right, then why this one.

novneet, your solution is impressive if the maximum number of splits is known. From what I understand, it needs to be dynamic.

Did a followup with my architect. He explained, as IBM put it, that we would require a java pack along with the webservice pack. Just wanted to let you know.

kommven wrote:You should be using ONLY ODBC Stage.

Psst.... DRS stage too

chulett wrote:select to_date(date1,'YYYY-MM-DD') from table

select to_date(date1,'YYYY-MM-DD HH24:MI:SS') from table
Two examples.

You mean select to_char (if its a date column)

As Craig requested earlier, please post your complete SQL.

For testing purposes, run the job with a row limit set to 1. See what happens.
What kind of update action do you have in the target OCI stage? Are you using generated sql or user defined?

Is this a recurring issue with this partuicular job?

On windows there is no .odbc.ini file. Refere to Ray's reply in thispost.

Well If you are using three different instances for loading data then I bet you will be using three different chunks of the entire data. You cannot load the same data file with all the instances. You will, just like the .ctl files, use three different .dat files, which will be parametrized as well.

TNS entries are only for Oracle and not for all databases. Do you mean the odbc connection info?