On Friday, May 31, 2002, at 02:13 , Sudarsan Raghavan wrote:
> while (<INPUTDATA>) {
> chomp;
> s/^\s+//;
> next if ((1 .. /^NPROC/) || m/^$/);
> # The conditions have been swapped here
> # Explanation for this is one of my earlier mails
> unless (/^Total/) {
> # Assumes the line to stop searching for input starts
> with Total
> my ($user, $mem, $cpu) = (split (/\s+/))[1, 4, 6];
> print "user = $user, mem = $mem, cpu = $cpu\n";
> }
> }
this is still generating:
user = USERNAME, mem = STATE, cpu = NICE
user = prago, mem = cpu49, cpu = 0
user = oracle, mem = cpu46, cpu = 0
user = prago, mem = sleep, cpu = 0
user = oracle, mem = sleep, cpu = 0
user = oracle, mem = sleep, cpu = 0
user = prago, mem = sleep, cpu = 0
user = USERNAME, mem = MEMORY, cpu = CPU
user = prago, mem = 2.9%, cpu = 11%
user = oracle, mem = 91%, cpu = 4.1%
user = patrol, mem = 0.2%, cpu = 0.2%
user = dbmsys, mem = 0.1%, cpu = 0.1%
user = root, mem = 0.8%, cpu = 0.1%
given the input data set of:
PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP
13261 prago 5728K 5240K cpu49 0 0 0:40.04 9.4% cfprsdrv/1
20318 oracle 519M 496M cpu46 0 0 35:47.20 3.2% oracle/1
12924 prago 1720K 1056K sleep 1 0 0:06.55 1.6% zcat/1
21244 oracle 514M 494M sleep 0 0 0:00.01 0.3% oracle/1
21107 oracle 526M 507M sleep 0 0 0:00.13 0.3% oracle/1
13310 prago 392M 101M sleep 59 0 0:01.07 0.2% syncsort/1
NPROC USERNAME SIZE RSS MEMORY TIME CPU
17 prago 1596M 394M 2.9% 0:48.54 11%
47 oracle 13G 12G 91% 38:17.25 4.1%
5 patrol 36M 23M 0.2% 10:53.17 0.2%
11 dbmsys 52M 20M 0.1% 0:00.53 0.1%
53 root 173M 113M 0.8% 5:32.40 0.1%
Total: 208 processes, 875 lwps, load averages: 2.03, 2.04, 2.12
cf:
http://www.wetware.com/drieux/pbl/Sys/Admin/ciscoParse2.txt
which given the same input file generates:
user prago hogs 2.9% memory and 11% cpu
user oracle hogs 91% memory and 4.1% cpu
user patrol hogs 0.2% memory and 0.2% cpu
user dbmsys hogs 0.1% memory and 0.1% cpu
user root hogs 0.8% memory and 0.1% cpu
we do not like:Total: 208 processes, 875 lwps, load averages: 2.03, 2.04,
2.12
which I think is closer to the OP's preferred choice.
and may I put a plug in for the
use Inline::Files;
WAY too wicked slick a solution for skanking test cases like this.
ciao
drieux
---
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
