How to manage Hash files? doing a new installation at client

ashcar · Post by **ashcar** » Wed Sep 29, 2004 9:12 am

I am installing DS7.5 on windows platform for a customer and had a question
In the Hash file stage under General-
we have option of using Account Name or Directory Path.. for storing Hash files.
What do you think is a better option to go with considering that the customer is new to DS and will be developing jobs for the first time.
from my earlier installation on Unix we were using directory path, so that we could delete hash files without touching the VOC and commands like create.file and delete.file.
i would appreciate ur comments as i never done installation on windows or worked on Datastage with windows

chucksmith · Post by **chucksmith** » Wed Sep 29, 2004 9:55 am

In currect releases, nothing is wrong with creating hash files in directories.

I strongly recommend your jobs ensure that the required hash files are created as part of the normal production process, not as part of an installation process. The easiest way to ensure the hash files exist is to simply link a transform stage to a hash file stage where the create file button is checked, and in the transform use a stage variable in the constraint that will always be false.

You can still create a VOC pointer to the file when required. There are many ways to do this, including my CreateVocPtr routine found on the tools page of www.anotheritco.com.

tonystark622 · Post by **tonystark622** » Wed Sep 29, 2004 10:57 am

I personally prefer to have hash files in a directory. There are already so many files in the project directory, when it comes time to try to go clean up the partition (you know you'll have to sometime) you're in danger of deleting files that you shouldn't or leaving files that can be deleted because you're not sure if they're DataStage files or not.

Good Luck,
Tony