Re: loop with awk

"Cameron Simpson [email protected] [sed-users]" <[email protected]> Sat, 3 Jun 2017 11:59:40 +1000
Newsgroups gmane.editors.sed.user
Message-ID <[email protected]>
On 03Jun2017 01:26, Daniel Goldman <[email protected]> wrote:
>"I've got a few remarks about your solution:" - That is very helpful to the 
>OP. FWIW, I prefer "folding style" #1. :) Just personal preference. Actually, 
>I would instead use temporary files, skip long pipelines. And speed is usually 
>a very secondary consideration, IMO. I think readability and maintainability 
>are usually much more important. "Avoid premature optimization". Anyway, it 
>was nice of you to help educate others.

Just regarding speed, with shell scripts the runtime is often dominated by the 
fork/exec cost because so many operations are farmed out to subprocesses.  
Because of this it is usually worth spending some thinking time on reducing the 
number of commands dispatched for the work (without making the script 
unreadable or unmaintainable). I agree about premature optimisation, but the 
threshold for _some_ fork/exec reduction is lower with the shell than in other 
languages.

Cheers,
Cameron Simpson <[email protected]>