Get output of fields starting from 2nd line

Shell Programming and Scripting
1

Hi All,

I am using the following command in Linux:

 
sar -r 30 3
Linux 2.6.18-194.3.1.7.3.el5xen  02/07/2013

02:55:47 PM kbmemfree kbmemused  %memused kbbuffers  kbcached kbswpfree kbswpused  %swpused  kbswpcad
02:56:17 PM 128646024  22348920     14.80    230232  15575860  75497464         0      0.00         0
02:56:47 PM 128646188  22348756     14.80    230244  15576116  75497464         0      0.00         0
02:57:17 PM 128646676  22348268     14.80    230248  15576160  75497464         0      0.00         0
Average:    128646296  22348648     14.80    230241  15576045  75497464         0      0.00         0
-bash-3.2$

I have tried using the following command to get the desired output but somehow not getting it. %memused field is blank

 
-bash-3.2$ sar -r 1 3 | grep -v Average | grep -v Linux | cut -d' ' -f1,2,6

12:57:57 PM %memused
12:57:58 PM
12:57:59 PM
12:58:00 PM
-bash-3.2$

From the above output, I need to get the following output onwards.

Output should look as below( Need only the values and not the columns to be displayed):

 
02:55:47 PM    %memused 
02:56:17 PM      14.80    
02:56:47 PM      14.80    
02:57:17 PM      14.80

Thanks for your time!

Regards,
a1_win

2 
sar -r 1 3 | awk '/^[0-9]/ {print $1,$2,$5}'
02:55:47 PM %memused
02:56:17 PM 14.80
02:56:47 PM 14.80
02:57:17 PM 14.80

PS remove the / in your start code tag

3

bash has a nice feature. You declare a simple array variable 'typedef -a myarray' and then 'while read -a myarray' and the fields are in the array for each line. Awk does that sort of thing natively.

BTW, free memory is a worthless spec in a VM system. That stat is left over from the old days of memory resident everything. delta page faults is more interesting, and load (cpu avg queue depth).

4

Instead of trying to filter away everything you do not want it is sometimes easier to specify what you do want: ever line where "<time> PM " is followed by a digit, not a letter, right?

grep '^[0-2][0-9]:[0-6][0-9]:[0-6][0-9] PM [0-9]'

Should filter all the lines you are interested in. As you use "cut" afterwards to dissect the lines you can combine this regexp and the "cut"ting into an awk- or sed-script, like Jotne already did - alas, his filter regexp is wrong:

sar -r 1 3 | awk '/^[0-2][0-9]:[0-6][0-9]:[0-6][0-9] PM [0-9]/ {print $1,$2,$5}'

I have to agree with DGPickett: you probably are extracting not particular meaningful numbers and might want to give your methodology a second thought.

I hope this helps.

bakunin

5

Thanks this has helped!

Regards,
a1_win