sed question

[Pete Des Moulins]

Hi experts,

Could someone please explain precisely what the following sed one line is doing :

bdf | sed '/:/N^J{^Js%\n% %^J}'

Thanks in advance and points will be given

[Pete Randall]

Not much:

# bdf | sed '/:/N^J{^Js%\n% %^J}'
sed: Function /:/N^J{^Js%\n% %^J} cannot be parsed.


Pete

[Sandman!]

It will join two adjacnt lines together. The "N" gets the next line of input and "s%\n% %" command replace the embedded newline with a space. I am not sure what "/:/" is doing. Is it giving an error on your machine?

~cheers

[Pete Des Moulins]

Thanks for the info, basically it currently does not work on our system, the end user is trying to prevent bdf output utilising 2 lines of output per file system.

One thought I had was to use bdfmegs ....

I'll know more on Monday.

[Tim Nelson · Accepted solution]

How about this instead.

bdf -l |awk '{if ( $2 == "" ) printf ;else print}'

[Pete Des Moulins]

Tim, that looks like a winner ;-)

[Sandman!]

>Thanks for the info, basically it currently does not work on our system, the end
>user is trying to prevent bdf output utilising 2 lines of output per file system.

In that case see the thread below:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1198814

[Dennis Handly]

This does what Sandman said except it only does it for lines with ":". So using the following would work nicely:
$ bdf -t nfs | sed '/:/N^J{^Js%\n% %^J}'

>Pete: Not much:

You need to convert ^J to newlines by inserting them with ^V^J. Or use:
$ bdf -t nfs | sed '/:/N
{
s%\n% %
}'