Table Conversion issue, DEV/TEST vs. PRODUCTION

jimmymac

Reputable Poster
We are on E1 9.0 with tools release 8.98.42.

We have many interfaces where we will process a text created by another system (peoplesoft), then we have an Interface in JDE to consume that file. We use Table Conversions in these processes to read data from the text file and write to the F0911Z1.

The predicament we have may be common and here it is. The incoming text files that we process come in with unique names, including a date/time stamp for example. The Table Conversion in JDE however needs a static name to be defined in the TC Design Aid, so we have a process that copies these text files as they are processed to a 'Workfile' with the same static name that is defined in the TC. Then the Table Conversion runs to process the data in the Workfile.

The problem comes in since this is a 'hard coded' file name in the Table Conversion, when we run a process in DEV or TEST, it uses the same workfile that will be used when this runs in production. So if we happen to be testing a Table Conversion, or doing QA testing that requires running a Table Conversion, at the same time that Table Conversion could be running in Production, there could be an issue where test data is pulled into production, or vice versa, or some other unpredictable results.

Does anyone have an approach where you can define a Table Conversion that has a different input file name and path when run in one environment, and another input file name and path when run in another environment.

Any thoughts would be appreciated.
 
Update. I may have found an alternative but want to see if anyone has used it and if there are any issues inherent in this.

What I see on Oracle Support is that in a Table Conversion, if you only specify a file name, not a full path, the default location of the file will be \pathcode\Import\.... So that way when run in production or development or test, the workfile we be used from the Import folder under whatever pathcode is being run at the time. So that way data from Production or Test could not cross contaminate.

Any one have experience with this solution?
 
Yes, our practice is to never supply a path for table conversions. Always use the default pathcode import/export location and employ a "wrapper" UBE to move the files to the locations needed (identfied by processing options) when reading or writing.

Craig
 
Back
Top