Hash file or indexed tables ?

hamzaqk · Post by **hamzaqk** » Tue Apr 17, 2007 5:58 am

Hi, working on DS 7.x . whats the need to use Hash files for joins when we can do the same with tables ? and if it is about hashing which makes it faster. in Teradata you can hash the key columns .... so when to choose what ? and why prefer hashfiles over indxed tables to carry out a join of the sequential file on the base table .

Thanks !

rkdatastage · Post by **rkdatastage** » Tue Apr 17, 2007 6:01 am

Hi
The advantage of using hashed file is that it will be stored in the server where the DS server is running . The turnaround time to read and write will be faster than reading/writing the data to database.
Added to this there are some more advantages using Hashed File.
Any one correct me if i am wrong .

Thanks
Ravi

DSguru2B · Post by **DSguru2B** » Tue Apr 17, 2007 6:52 am

Also takes the load off your database server.

ray.wurlod · Post by **ray.wurlod** » Tue Apr 17, 2007 2:21 pm

It's possibly even better than being local; DataStage hashed files can be cached in memory, so that the fast key-based access (via the hashing algorithm) is at least three orders of magnitude faster than disk-based access.

In DataStage the term is "hashed" file, not "hash" file.