Odtnhj

I have the following:

2018-11-10 23:57:21 [COMMAND]: sar -u 10 5

AIX host 1 7 11/10/18

System configuration: lcpu=64 mode=Capped

23:57:21 %usr %sys %wio %idle physc
23:57:31 10 7 0 83 16.00
23:57:41 9 6 0 85 16.00
23:57:51 9 6 0 85 16.00
23:58:01 9 7 0 84 16.00
23:58:11 10 6 0 84 16.00

Average 9 6 0 84 16.00

2018-11-10 23:58:21 [COMMAND]: sar -u 10 5

AIX host 1 7 11/10/18

System configuration: lcpu=64 mode=Capped

23:58:21 %usr %sys %wio %idle physc
23:58:31 10 8 0 82 15.99
23:58:41 9 6 0 85 16.00
23:58:51 9 6 0 85 16.00
23:59:01 9 6 0 84 16.00
23:59:11 10 6 0 83 16.00

Average 10 6 0 84 16.00

I need to get the time with the average value of %idle:

2018-11-10 23:57:21|84
2018-11-10 23:58:21|84

Going by your input file as-is, a simple awk command as simple as below should suffice.

awk '/sar/ time=$1" "$2; next /Average/"$5 ' file

A couple of choices:

• 
  

  awk

  
  
  

  $ awk '/COMMAND/printf "%s %s", $1,$2/Average/print "' file 
  2018-11-10 23:57:21|84
  2018-11-10 23:58:21|84
  

  
  


• 
  

  perl

  
  
  

  $ perl -ale '$k="$F[0] $F[1]" if /COMMAND/; print "$k|$F[4]" if /Average/' file 
  2018-11-10 23:57:21|84
  2018-11-10 23:58:21|84
  

  
  

This Perl one-liner should do the trick:

perl -lne'$d=$1 if /^(d4-d2-d2s+d2:d2:d2)/;
if(/^Averages+S+s+S+s+S+s+(S+)/) print "$d'

(I made it two lines, so it's more readable.) You need to feed it with input though. The file name at the end, or <file_name at the beginning or the end, or a pipe will do. The workings are these:

1. Go to a line with date and time at the beginning and memorise it.


2. Go to a line with Average at beginning and once there, print the memorised date and time plus the | sign and the average taken from the fifth column assuming whitespace column separation.


3. Point 1. repeats.