BASH: Find number in text -> variable

I need little help from the community:

I have these two lines in a large text file:

Connected clients: 42  
4 ACTIVE CLIENTS IN LAST 20 SECONDS  

How I ca开发者_高级运维n find, extract and assign the numbers to variables?

clients=42
active=4

SED, AWK, GREP? Which one should I use?

---

clients=$(grep -Po '^(?<=Connected clients: )([0-9]+)$' filename)
active=$(grep -Po '^([0-9]+)(?= ACTIVE CLIENTS IN LAST [0-9]+ SECONDS$)' filename)

or

clients=$(sed -n 's/^Connected clients: \([0-9]\+\)$/\1/p' filename)
active=$(sed -n 's/^\([0-9]\+\) ACTIVE CLIENTS IN LAST [0-9]\+ SECONDS$/\1/p' filename)

---

str='Connected clients: 42 4 ACTIVE CLIENTS IN LAST 20 SECONDS'

set -- $str
clients=$3
active=$4

---

If it's two lines, fine.

str1='Connected clients: 42'
str2='4 ACTIVE CLIENTS IN LAST 20 SECONDS'

set -- $str1
clients=$3
set -- $str2
active=$1

Reading two lines from a file may be done by

{ read str1; read str2; } < file

---

Alternately, do the reading and writing in AWK, and slurp the results into Bash.

eval "$(awk '/^Connected clients: / { print "clients=" $3 }
             /[0-9]+ ACTIVE CLIENTS/ { print "active=" $1 }
            ' filename)"

---

you can use awk

$ set -- $(awk '/Connected/{c=$NF}/ACTIVE/{a=$1}END{print c,a}' file)
$ echo $1
42
$ echo $2
4

assign $1, $2 to appropriate variable names as desired

if you can directly assign using declare

$ declare $(awk '/Connected/{c=$NF}/ACTIVE/{a=$1}END{print "client="c;print "active="a}' file)
$ echo $client
42
$ echo $active
4