Re: [1/2OT] something about sed

[lina <lina.lastname@gmail.com>]

On Monday 27,February,2012 12:14 AM, Javier Barroso wrote:
> On Sun, Feb 26, 2012 at 4:30 PM, lina<lina.lastname@gmail.com>  wrote:
> > Hi,
> >
> > sorry a bit off-topic,
> >
> > sed -e 's/\(<[^ ]*>\)\([ ]*\)\(<[^ ]*>\)/\3\2\1/g'
> > GNU Linux is cool
> > Linux GNU cool is
> No here ...
>
> > I don't know why
> >
> > $ echo "GNU linux is cool" | sed -e 's/\(<[^ ]*>\)\([ ]*\)\(<[^ ]*>\)/\3\2\1/g'
> > GNU linux is cool
> >
> > It doesn't work.
>
> Remove "<" and">", from substitution, and sustitute * by "\+", and
> you can use " *" without have to use "[ ]*"

I read it from a book and didn't get why use "<"

> sed -e 's/\([^ ]\+\)\( \+\)\([^ ]\+\)/\3\2\1/g'

Thanks, I tried, sed -e 's/\([^ ]\+\) \([^ ]\+\) \([^ ]\+\)/\3 \2 \1/g'

> or using extended regular expressions, you can remove all escape stuff :
>
> sed -re 's/([^ ]+)( +)([^ ]+)/\3\2\1/g'

Ha ... I didn't realize the -r can be used.

sed -re 's/([^ ]+) ([^ ]+) ([^ ]+)/\3 \2 \1/g'

> I hope this help

Thanks again for your help. I was a bit lost when reading other's work, 
... now better. ^_^
>

Best regards,