DSXchange

It may be even more than 320 * 8 connections opened because if your jobs have more than one Oracle EE stage e.g an input stage and and output stage then that's 2 connections per job. Best to scale back number of jobs that you run at the same time.

perhaps you could provide more details on what you're trying to achieve and clearly define integrate?
i guess the underlying metadata has to be consistent for any integration work.you can't integrate when both the data marts speak different languages if you know what i mean.

You can still download evaluation version but you must contact IBM for the evaluation license key.

If you require further information please contact Sales Administration.
http://www.ascential.com/iisp/public/contact.do

Product downloads can be found at:
https://www14.software.ibm.com/webapp/i ... e=wiiseval

You have a data quality issue here, the source data contains row(s) with the value "WS " which you are attempting to convert to a decimal. either you provide rules to deal with this e.g what is the equivalent of WS in decimal ? or you filter out these records.

[quote] My question is: how can the performance be improved?

hire a consultant to help you tune your jobs?

firstly i would suggest downloading thesoftware off the IBM site. doing an install from a tar file is not the right thing to do. you might get it to work but that is not a proper installation and is fraught with traps. if you already have the license keys,then the installing the downloaded copy shou...

you should get a baseline first in how long it takes to sort by sorting that file outside of ds, try just using the good old fashioned unix sort and see how long that takes. if your unix sort too takes 4 hours or more then the problem is probably due to your environment (hardware,cpu,resources). if ...

perhaps a better way would be to tell us what you want to do? do you intend to do a lot of custom sql ? if that is the case then API for everything else i think EE would suffice.

interesting that projects move from dev to prod without test

you'll prob need a sequencer to get the start and end time of the job cos the end time of a job is only known after job has finished.
try doing a search on etlstats.

John Smith, Can you give a few notes about the both choices as your say? typically DS is run in client server env.by two tier i mean you have clients and a server running the dsrpc process.client submit jobs to server and jobs are run in the server. three tier would be like a mainframe env where ds...

try just bumping it up to 100 and see if it helps. it's just a suggestion. you can do that within the job itself.

srikanth386 wrote:wht is the architecture of datastage ?

client server? two tier or three tier? take your pick.

to start with why is your array size set to 1 ? that is very low which means you are commiting every row.try increasing the array size if you can and you will see some improvements.