Virtual Hash files , Distributed Hash file

Puli · Post by **Puli** » Tue Jun 10, 2008 11:38 pm

Hi All,

Last day i came across 2 new interesting words.One is Virtual Hash file and the other is Distributed Hash file. You people have any idea on these or are they some scrap idea?

Give your thoughts..

Regards
Puli

ray.wurlod · Post by **ray.wurlod** » Wed Jun 11, 2008 12:46 am

Where did you come across these terms?

My guess is that a virtual hashed file (never "hash" file) is one that is cached in memory.

A Distributed File is a UniVerse construct that is also available in DataStage; it is a logical way to provide an "umbrella" containing multiple standard hashed files that have identical metadata. Before 64-bit addressing became available in standard hashed files, Distributed Files were one way to defeat the 2GB size limit on hashed files.

Puli · Post by **Puli** » Wed Jun 11, 2008 4:09 am

Hi,

Actually last week we had a discussion on Hash file design considerations and performance tuning in hash file. So during that one delegate was giving his comments on this.I guess the Virtual concept is similar to the Cached hased file..but distributed...i dont remember the scenario in which he presented that word...

Regards
Puli

ArndW · Post by **ArndW** » Wed Jun 11, 2008 4:17 am

Maybe this person was thinking of the PX virtual datasets and applied it (incorrectly) to server hashed files.

chulett · Post by **chulett** » Wed Jun 11, 2008 6:34 am

I was wondering what the answer would be to this. I for one have never heard of a 'virtual' hashed file but doesn't mean much so waited for an official Wurlod answer to see.

Distributed hashed files, however, are a Known Thing and something I've experimented with a number of years ago after seeing an interesting presentation on their usage (by BCBS, I think) at an AscentialWorld back in the day.