[SOLVED] Need help in displaying particular lines via grep

khandu · 04-08-2011, 07:09 AM

So For example I run this command

df -h

I get

Code:

Filesystem      Size   Used  Avail Capacity  Mounted on
/dev/disk0s2   373Gi  271Gi  102Gi    73%    /
devfs          178Ki  178Ki    0Bi   100%    /dev
map -hosts       0Bi    0Bi    0Bi   100%    /net

But want to gerp / cut it in such a way that it only displays

Code:

/dev/disk0s2   373Gi  271Gi  102Gi    73%    /
devfs          178Ki  178Ki    0Bi   100%    /dev
map -hosts       0Bi    0Bi    0Bi   100%    /net

Now the thing is that these 3 lines are not static.. there can be N number of lines there.. the only thing is that I want the command / output NOT to display the first line but the rest of the n lines .. how can I do this?

thanks

syg00 · 04-08-2011, 07:15 AM

What have you tried - what didn't work ?.

khandu · 04-08-2011, 07:18 AM

Well

I tried a tail but then that will get stuck with limited lines

So if I do a df -h | tail -3

I will get that 3 lines.. But if I have more mounts.. they will not show.. and only the last three lines will..

syg00 · 04-08-2011, 07:31 AM

If you always want lines 2 till the end, try sed - it has an option for lines ranges that includes a non-specific descriptor for the last line.
awk or perl can also handle it easily.

khandu · 04-08-2011, 07:35 AM

I am quite a n00b.. Can u show how to do it? I am not sure..

syg00 · 04-08-2011, 07:46 AM

Start here.
The sed one-liners is a (very) handy reference.

khandu · 04-08-2011, 07:53 AM

Quote:

Originally Posted by syg00

Start here.
The sed one-liners is a (very) handy reference.

Awesome

so this works

Code:

df -h | sed '/Size/d'

Any better / more elegent way to do it? or is this good?

syg00 · 04-08-2011, 08:00 AM

Many ways to skin a cat.
Whatever works - I'd suggest you also try variants using line (number) ranges.

MTK358 · 04-08-2011, 08:42 AM

To remove the first line:

Code:

df -h | sed '1d'

colucix · 04-08-2011, 09:03 AM

Using tail you can specify the line to start printing till the end of the file:

Code:

df -h | tail -n +2

Just to add an awk solution as per "many ways to skin a cat" cited above:

Code:

df -h | awk 'NR > 1'

An aside note: if you want to process the output of the df command in a script, better to use the -P (portability) option. It ensures that the output is not splitted into multiple lines, for example when the device name is long. For example on a machine having logical volumes with default names:

Code:

$ df -h
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/isw_dcjhjdcjac_Volume0p2
                      379G   11G  348G   4% /
/dev/mapper/VolGroup00-LogVol00
                      1.8T  212G  1.5T  13% /home
/dev/mapper/isw_dcjhjdcjac_Volume0p1
                      114M   24M   85M  22% /boot
tmpfs                 5.9G     0  5.9G   0% /dev/shm
/dev/sdi1             917G  170G  701G  20% /media/backup_op_004
$ df -Ph
Filesystem            Size  Used Avail Use% Mounted on
/dev/mapper/isw_dcjhjdcjac_Volume0p2  379G   11G  348G   4% /
/dev/mapper/VolGroup00-LogVol00  1.8T  212G  1.5T  13% /home
/dev/mapper/isw_dcjhjdcjac_Volume0p1  114M   24M   85M  22% /boot
tmpfs                 5.9G     0  5.9G   0% /dev/shm
/dev/sdi1             917G  170G  701G  20% /media/backup_op_004

The second one is less human readable, but more manageable from a script. Just my