handling lookup data

just4u_sharath · Post by **just4u_sharath** » Sat Nov 03, 2007 8:06 pm

i have a question. A lookup stage in parallel edition can accomodate how much amount of data. becuase i want to differentiate between a join and a lookup. i heard for less data we use lookups and for huge datas we use join stage. so i want to know till what amount of memory(in bytes) i can user a lookup

ArndW · Post by **ArndW** » Sat Nov 03, 2007 8:20 pm

It is not absolute. It depends upon your available memory, lookup source type, speed, performance expections, levels of parallelism and other factors.

just4u_sharath · Post by **just4u_sharath** » Sat Nov 03, 2007 9:08 pm

in our case we are running on two nodes and each node has a memory on 1 MB. with this info can i get the answer i require

ray.wurlod · Post by **ray.wurlod** » Sat Nov 03, 2007 10:38 pm

With 1MB of memory you're not even going to have an operating system, much less DataStage. Please check your figures.

just4u_sharath · Post by **just4u_sharath** » Sat Nov 03, 2007 11:24 pm

ray.wurlod wrote:With 1MB of memory you're not even going to have an operating system, much less DataStage. Please check your figures.

i am sorry each node carries 1 GB of memory

ArndW · Post by **ArndW** » Sun Nov 04, 2007 12:50 am

A "node" is a PX/EE term and you cannot directly associate memory to it.

Do you mean you have a 2-CPU system with a total of 2Gb main memory?