Hash file in account/directory path

saprebv · Post by **saprebv** » Tue Oct 26, 2004 4:09 am

Is there any disadvantages/advantage of creating hash file in account over creating hash file in account?

Thank you,
Regards,
Balkrishna Sapre

saprebv · Post by **saprebv** » Tue Oct 26, 2004 4:10 am

sorry .
read question as

Is there any disadvantages/advantage of creating hash file in account over creating hash file in directory path?

jeredleo · Post by **jeredleo** » Tue Oct 26, 2004 6:15 am

Wouldn't mind knowing this info myself. We use the directory path mainly due to the lack of Universe commands to maintain the hash files through the administrator. This way we can copy and move our hash files to where-ever we need them to be.

JB

kduke · Post by **kduke** » Tue Oct 26, 2004 4:37 pm

JB is correct. You can move files around because the VOC in the account is not updated. You can still move them if you know Universe but you may have to switch to a file path. Having a VOC entry allows you to LIST or SQL SELECT against your hash file file. It is also easier to run HASH.HELP or ANALYZE.FILE. This can be overcome as well with SETFILE.

I have no personal preference. I think both are fairly easy to use. I think directory paths are a little more maintenence. If you create one big filesystem with all your accounts in it then it is easier to use account. If you have lots of large files on several filesystems then use path. The larger the files the more likely you will need to use paths and learn to administor Universe.

ray.wurlod · Post by **ray.wurlod** » Tue Oct 26, 2004 8:56 pm

I also have no great preference. In my projects I generally arrange for the hashed files to be in a directory (or sometimes more than one, functionally grouped). This means that the project directory itself doesn't grow out of control!

However, I also arrange for VOC pointers to exist so that UniVerse adminstration tools can be used. This can be done with SETFILE or by creating the hashed files with CREATE TABLE with its DATA and DICT extensions (search forum for more on this).

If the hashed file or directory is ever moved, this approach adds the miniscule overhead of needing to update the VOC pointer (using SETFILE with the OVERWRITING keyword).