Substitute 2 character country abbreviations to 3

hattori.hanzo · 12-20-2010, 10:57 PM

I have a list of 2 character country abbreviations used in the Maxmind GeoCountry database that I need to convert to the standard 3 character country abbreviation ISO 3166-11 format.

I could have a long list of sed substitute statements for each country.

Code:

sed "s/US/USA/g" <file>

Is their a more elegant way to do this? Maybe an array of some sort?

This substitution will be used in a bash script.

Thanks & Regards,

David the H. · 12-21-2010, 04:30 AM

If we could see an example of the full input format we could probably come up with some other options. A loop and/or case statement using bash parameter substitution, for example, or an awk script.

At the very least, you could create a script of sed substitution expressions, instead of running it as a series of individual commands.

In the end though, each of the possible codes will probably have to be set up individually somehow.

grail · 12-21-2010, 05:28 AM

I would suggest having a file with the change you require and then looping over it:

Code:

US USA
AU AUS
...

Also be wary of your sed as it would be most unfortunate to get something like:

Code:

USASR

If USSR were somewhere in the file

hattori.hanzo · 12-22-2010, 11:03 PM

Thanks for the pointer.

Here is the test data:

Code:

[me@host test]$ more country.list
HK|HKG
US|USA
AU|AUS
[me@host test]$ more country.tmp
HK
US
JP
AU

Using awk from an example here.

Code:

[me@host test]$ awk 'BEGIN {FS=OFS="|"} FNR==NR{a[$1]=$2;next} $1 in a{print $2}' country.tmp country.list

Output

Code:

HKG
USA
AUS

If a country abbreviation is in country.tmp and not in country.list how could I display 'unknown' for that entry?

Update

Example taken from here which seems to work.

Code:

cat country.tmp | while read num; do NAME=`awk -F "|" '$1=="'"$num"'" {print $2}' country.list`; [ -z $NAME ] && echo "unknown" || echo $NAME; done

Output

Code:

HKG
USA
unknown
AUS

Thanks & Regards

grail · 12-23-2010, 12:10 AM

Well I would probably stick with one or the other, ie either awk or bash.

awk

Code:

awk 'BEGIN{FS="|"}FNR==NR{a[$1]=$2;next}{if($1 in a)print a[$1];else print "unknown"}' country.tmp country.list

bash

Code:

declare -A arr

while IFS="|" read -r n2 n3
do
    arr[$n2]=$n3
done<country.tmp

while read -r line
do
    [[ -n ${arr[$line]} ]] && echo ${arr[$line]} || echo unknown
done<country.list