Re: finding .m files using 'locate'

[Peter J. Acklam]

Mike Miller wrote:
>
> Peter J. Acklam wrote:
>
> > The -l (ell) option is very useful for one-liners like this
> >
> >  locate "*.m" | perl -ple 's#(^.*)/[^/]+$#$1#' | sort | uniq
> >
> > which can be simplified to
> >
> >  locate "*.m" | perl -ple 's|/[^/]+$||' | sort | uniq
> 
> Shoter *and* slightly faster!

That's because in this case perl doesn't have to store the captured info into 
the $1 variable and insert it afterwards.

> > where the "sort" and "uniq" can be removed by using
> >
> >  locate "*.m" | perl -nle 's|/[^/]+$|| ; print unless $d{$_}++'
> 
> Slightly faster still, but also a little longer.

It also saves a few processes. You can micro-optimize the above into

... | perl -nle 's|/[^/]+$||; $d{$_}++ || print'

Peter