Hi,
I've written a one line script to consolidate ping stats for each device being pinged from a list of 30 devices. I am using a for-do loop and extracting and formatting the values. But I'm having problems extracting just the average value:

My output so far looks like this:

MyCity to device.domain.com (10.10.10.12): 04-30-08 13:05:00 min/avg/max = 103/106/113

I need it to show only the average value, like this:
MyCity to device.domain.com (10.10.10.12): 04-30-08 13:05:00 avg = 106

How can I accomplist this since the value average is not always 3 characters? I tried using nawk and sed, but no luck yet extracting just the average for each ping result.

Thanks

Consider this:

echo "MyCity to device.domain.com (10.10.10.12): 04-30-08 13:05:00 min/avg/max = 103/106/113" | nawk
' BEGIN { FS="/" } { $3="="
$5=""
gsub("min","")
print $0} '

It's a kludge and it's easily broken if the data structure varies from what you've said.

Using a field seperator, FS, equal "/", then
field 3 should be replaced by an equal sign and field 5 should be nulled out. That leaves the "min" string which is removed by a call to gsub.

awk '{
 n=split($NF,a,"/") 
 for(i=1;i<=n;i++) total+=a[i]
 $NF = total/n
 print 
}
' output

ghostdog:

I executed your awk script and this is the output I received:

MyCity to device.domain.com (10.10.10.12): 04-30-08 13:05:00 min/avg/max = 107.333

The requirement is for the end of the string should look like this:

... avg = 106

nails:

maybe i misinterpreted what OP wants. I thought he meant to take the average of the 3 numbers. So its the middle number he wants..therefore, the for loop should be removed, and a[2] will be the average nnumber... thanks for pointing out

Thanks all for the suggestions. Not the cleanest, but it did the job for me:
#!/bin/ksh

DATE=`date '+%m-%d-%y %H:%M'`

for devices in `cat /tmp/devices` do
/usr/sbin/ping -ns $devices 56 2 | sed -e '2,5d' -e 's/PING /From City to/' -e 's/56 data bytes//' -n -e '/(/p' | sed -e :a -e '$!N;s/\nround-trip (ms) min\/avg\/max =/'"$DATE"'/' |
nawk -F/ '{print $1, $2}' | nawk '{print $1"," $2"," $3"," $4"," $5"," $6}' >> /tmp/results