Search code examples
linuxbashawksh

Define variable in AWK

I am trying to mention varaibel in AWK. I try a few I used to use earlier, but can't figureout where I am doing wrong...

cat tld.csv
co.uk
com

code

while read -r ip 
do

awk -v i="$ip" -F, '$1 ~ /i$/ {print $0}' full-database.csv > tld-specific-database/$ip.csv


done < tld.csv

cat full-database.csv

a.com,bbb
co.uk,ddd
a.comi,bbb
c.ukip,ddd

Expected result

cat co.uk.csv
    co.uk,ddd

cat com.csv
    a.com,bbb

what i am getting now is

cat co.uk.csv
    a.comi,bbb

cat com.csv
    a.comi,bbb
Solution

  • What you want to match is the value of variable i, not the actual character (or string) i:

    awk -F, -v i="$ip" '$1 ~ i "$"' db.csv >$ip.csv

    or slightly terser include the $ in the variable value

    awk -F, -v i="$ip$" '$1 ~ i' db.csv >$ip.csv

    (either way the default action is print $0 so you can leave that out).

    But you can do this in one pass almost as https://stackoverflow.com/users/3422102/david-c-rankin says, if the number of patterns in tld.csv is reasonably small as you show, and this is more efficient if fulldb.csv is large (which you don't show, but 'full' suggests):

    awk -F, 'NR==FNR{x[$0];next} { for(i in x) if($1 ~ i "$") print > i ".csv" }' tld.csv fulldb.csv

    or if you don't like the NR==FNR trick

    awk -F, 'BEGIN{while((getline<"tld.csv")>0) x[$0]} { for(i in x) ... }' fulldb.csv