processing

kris_daredevil · Post by **kris_daredevil** » Wed Oct 18, 2006 9:47 am

Hi All,

Is it good practice to do transformation into one variable
e.g trim(ereplace(DSLINK.col1," ","-")
or do
trim(DSLINK.col1)=variable1
ereplace(variable1," ","-") =variable2

how does the process work in both the cases

thanks
kris

kcbland · Post by **kcbland** » Wed Oct 18, 2006 9:54 am

Ignoring that your examples give different results (trim and ereplace vs. ereplace and trim), it's perfectly fine to nest functions, just keep your logic legible. Consider using stage variables when the same derivation of a value is repeated for many columns, or you need to simply a derivation/constraint.

ArndW · Post by **ArndW** » Wed Oct 18, 2006 9:59 am

The first one performs only one variable assignment, the second one needs to do 2; so the first will be more efficient. As Ken has stated, they will in their current form give different results.

DeepakCorning · Post by **DeepakCorning** » Wed Oct 18, 2006 10:09 am

Your first scenario does not match with the second one, so my guess is that you are talking about nesting many functions tgether and performace wise which one is better? I think nesting is perfectly well and performance wise I have not seen any jobs which perform bad if they have nested function.

kris_daredevil · Post by **kris_daredevil** » Wed Oct 18, 2006 10:36 am

Thanks All,

Assuming both the results give the same output .
iam sorry for posting wrong way.
iam just intrested in knowing how the internal process works is their any kind of documentation links or .pdf where i can know how the datastage processes the incomming data with the stages.

thanks once again for replies..

kris

meena · Post by **meena** » Wed Oct 18, 2006 1:02 pm

Hi Kris,
There may be many ways to write a code.But need to check which is the fastest.Mean to consider the performace of a job.And Online documentation or .pdf files are available when you install the Datastage on your system.