Using AWK for best match replace - awk

Question

I have two files:
operators.txt # includes Country_code and Country_name
49 Germany
43 Austria
32 Belgium
33 France
traffic.txt # MSISDN and VLR_address (includes Country_code prefix)
123456789 491234567
123456788 432569874
123456787 333256987
123456789 431238523
I need to replace the VLR_address in traffic.txt file with Country_name from the first file.
The following awk command do that:
awk 'NR==FNR{a[$1]=$2;next} {print $1,a[$2]}' <(cat operators.txt) <(cat traffic.txt|awk '{print $1,substr($2,1,2)}')
123456789 Germany
123456788 Austria
123456787 France
123456789 Austria
but how to do it in case operators file is:
49 Germany
43 Austria
32 Belgium
33 France
355 Albania
1246 Barbados
1 USA
when country_code is not fixed length and in some case best match will apply e.g.
124612345 shall be Barbados
122018523 shall be USA

Answer 1

The sample input/output you provided isn't adequate to test with as it doesn't include the cases you later described as problematic but if we modify it to include a representation of those later statements:
$ head operators.txt traffic.txt
==> operators.txt <==
49 Germany
43 Austria
32 Belgium
33 France
1 USA
355 Albania
1246 Barbados
==> traffic.txt <==
123456789 491234567
123456788 432569874
123456787 333256987
123456789 431238523
foo 124612345
bar 122018523
then this may be what you want:
$ cat tst.sh
#!/usr/bin/env bash
awk '
NR==FNR {
keys[++numKeys] = $1
map[$1] = $2
next
}
{
for (keyNr=1; keyNr<=numKeys; keyNr++) {
key = keys[keyNr]
if ( index($2,key) == 1 ) {
$2 = map[key]
break
}
}
print
}
' <(sort -k1,1rn operators.txt) traffic.txt
$ ./tst.sh
123456789 Germany
123456788 Austria
123456787 France
123456789 Austria
foo Barbados
bar USA

Answer 2

You obviously need to try a substring of the correct length.
awk 'NR==FNR{a[$2]=$1;next}
{ for (prefix in a) {
p = a[prefix]; l = length(p)
if ($2 ~ "^" p) { $2 = prefix; break } } }1' operators.txt traffic.txt
Notice how Awk itself is perfectly capable of reading files without the help of cat. You also nearly never need to pipe one Awk script into another; just refactor to put all the logic in one script.
I swapped the value of the key and the value in the NR==FNR block but that is more of a stylistic change.
And, as always, the final 1 is a shorthand idiom for printing all lines.
Perhaps as an optimization, pull the prefixes into a regular expression so that you can simply match on them all in one go, instead of looping over them.
awk 'NR==FNR{a[$1]=$2; regex = regex "|" $1; next}
FNR == 1 { regex = "^(" substr(regex, 2) ")" } # trim first "|"
match($2, regex) { $2 = a[substr($2, 1, RLENGTH)] } 1' operators.txt traffic.txt
The use of match() to pull out the length of the matched substring is arguably a complication; I wish Awk would provide this information for a normal regex match without the use of a separate dedicated function.