tail -f helping

dr.x · 10-05-2019, 10:27 AM

tail -f /var/log/messages/*-sockd.log | grep --line-buffered google | awk '{print $1,$6,$12,$17}'

05/Oct/2019:17:18:57 23.60.15.126 jan 515607
05/Oct/2019:17:18:57 2.147.234.142 minh 3342
05/Oct/2019:17:18:57 2.147.235.85 minh 442
05/Oct/2019:17:18:57 2.147.11.102 minh 333

but i want to count the last variable ,
i tried
tail -f /var/log/messages/*-sockd.log | grep --line-buffered google | awk '{print $1,$6,$12,$17}' | wc -m $17

but it fail , how can i count the digits in the output of the $17 variable ?

Thanks

teckk · 10-05-2019, 12:03 PM

Code:

list=(
'05/Oct/2019:17:18:57 23.60.15.126 jan 515607'
'05/Oct/2019:17:18:57 2.147.234.142 minh 3342'
'05/Oct/2019:17:18:57 2.147.235.85 minh 442'
'05/Oct/2019:17:18:57 2.147.11.102 minh 333'
)

for i in "${list[@]}"; do
    echo "${#i}"
done

44
44
42
42

teckk · 10-05-2019, 12:14 PM

Code:

strng='05/Oct/2019:17:18:57 2.147.11.102 minh 333'

wc <<< "$strng"
1  4 43
#1 line, 4 words 43 chars.

echo "${strng//[!0-9]/}"
#Just the intergers

echo "${strng//[!a-z]/}"
#Just the chars

That last one missed the capital O in Oct. I'll let you do that one.

pan64 · 10-05-2019, 12:19 PM

$17 exists only inside the awk script (which is now a single print command).
If you want to count you need to do it inside the awk script.
Actually grep is superfluous, you can do something like:

Code:

awk '/google/ { print $1,$6,$12,$17 }' /var/log/.....

and also you can make some other calculations instead of print.
But tail -f will be never finished, so your calculation will never be ready.

Turbocapitalist · 10-05-2019, 12:19 PM

I guess you could count the decimal places.

Code:

... | awk '/somepattern/{s=$17;d=0;while(s){s=int(s/10);d++};print $1,$6,$12,$17,d;}'

sub() could work that way, instead of int().