regex - Bash, Netcat, Pipes, perl -

background: have simple bash script i'm using generate csv log file. part of bash script poll other devices on network using netcat. netcat command returns stream of information can pipe grep command values need in csv file. save return value grep bash variable , @ end of script, write out saved bash variables csv file. (simple enough.)

the change i'd make amount of netcat commands have issue each piece of information want save off. each issued netcat command possible values returned (so each time returns same data , burdensome on network). so, i'd use netcat once , parse return value many times need create bash variables can later concatenated single record in csv file i'm creating.

specific question: using bash syntax if pass output of netcat command file using > (versus current grepmethod) file each entry on own line (presumably separated \n eol record separator -- easy perl regex). however, if save output of netcat directly bash variable, , echo variable, of data jumbled together, cumbersome parse out (not easy).

i have played 2 options: first, think perl one-liner may solution here, i'm not sure how best execute it. pseudo code might save netcat output a bash variable , somehow figure out how parse perl (not straight forward though).

the second option use bash's > , send netcat's output file. easy process perl , regex given \n eol, require opening external file , passing perl script processing , somehow passing return value bash script bash variable entry csv file.

i know i'm missing simple here. there way can force newline entry bash variable netcat , repeatedly run perl-one liner against variable create each of csv variables need -- within same bash script? sorry, long question.

the second option use bash's > , send netcat's output file. easy process perl , regex given \n eol, require opening external file , passing perl script processing , somehow passing return value bash script bash variable entry csv file.

this common idiom: save output netcat in temporary file, use grep or awk or perl or what-have-you many times necessary extract data file:

# create temporary file , arrange have  # deleted when script exists. tmpfile=$(mktemp tmpxxxxxx) trap "rm -f $tmpfile" exit  # dump data netcat # temporary file. nc somehost someport > $tmpfile  # extract information variable `myvar` myvar=$(awk '/something/ {print $4}' $tmpfile) 

that last line demonstrates how output of (in case, awk script) variable. if using perl extract information same thing.

you write whole script in perl, might make life easier.