Link Partitioner Hashing Algorithm Question

[jpr196]

Hi,

We have multiple jobs with large volumes of data where we've partitioned the jobs to load from a single DB2 source query into multiple DB2 target stages...though same table. Something like this (with some other processing in between):


|---------->tgt
Src---->Part------->tgt
|---------->tgt

Anyway, these tables are partitioned using DB2's hash function. Right now, we're using round-robin to partition our data, but the dba's suggested it might be quicker if we can break up our data streams to insert into each table partition. My question is, does anyone know if the Ascential Hash Algorithm works similarly to the DB2 hash algorithm or would we be wasting our time changing these link partitioner stages to use the Hash Algorithm instead of round-robin? Thanks in advance!

Edit...my little drawing isn't coming out how I wanted it to, but the pipes should be coming out of the part stage not the source.

[chulett]

Use 'code' tags to preserve whitespace and liberal use of the Preview function combined with editing to adjust for the proportional font being used:

Code: Select all

         |---------->tgt
Src---->Part-------->tgt
         |---------->tgt

[ray.wurlod]

You can wrap your diagram in Code tags.

The hashing algorithm used in the link partitioner is the GENERAL hashing algorithm of dynamic hashed files. It is unrelated to the DB2 hashing algorithm.

The answer to your next question, about whether this can be customized, is "no".

[jpr196]

I'll remember the code tags for next time. Thanks alot for the response...that's what I kind of figured regarding the hash algorithm.