Convert ISO Date to DD-MON-YYYY in a csv file in unix - awk

Question

I've csv file containing date in ISO format like below.
id,x1,x2,x3
AIR,Partner,2015-10-20T04:00:00.000Z,2015-10-20T04:00:00.000Z,2016-02-12T05:00:00.000Z
CMX,Partner,Tier,2017-03-23T04:00:00.000Z
WKA,Partner,Tier,2017-05-22T04:00:00.000Z
APP,Partner,2017-10-04T04:00:00.000Z,Tier
BUN,2017-09-27T04:00:00.000Z,Partner,,2017-09-27T04:00:00.000Z
There is no fix column for date it can appear in any column except 1st column.
I want to convert all occurrence of ISO date into DD-MON-YYYY or DD/MM/YYYY format.
Please help.

Answer 1

You can use the following through a UNIX pipe:
sed -E 's#\b([0-9]{4})-([0-9]{2})-([0-9]{2})T[0-9]{2}:[0-9]{2}:[0-9]{2}.[0-9]{3}Z\b#\3/\2/\1#g'
Another option would be the following PCRE, which explicitly enforces the , separator at the start (as you mention this field cannot be the first one), and either , or EOL at the end of the matched expression:
cat | perl -pe 's#,\K([0-9]{4})-([0-9]{2})-([0-9]{2})T[0-9]{2}:[0-9]{2}:[0-9]{2}.[0-9]{3}Z(?=,|$)#$3/$2/$1#g'

Answer 2

You can combine the previous answer with How to Print Spelled out month names rather than numbers to get the following two options, depending on which of the two answers you like best:
The first option is to not use any libraries, but just use the direct embedding of the names of the months:
cat 48941818.txt | perl -pe '#m = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec); s#,\K([0-9]{4})-([01][0-9])-([0-3][0-9])T[0-9]{2}:[0-9]{2}:[0-9]{2}.[0-9]{3}Z(?=,|$)#$3-$m[$2]-$1#g'
The second option is to use http://perldoc.perl.org/POSIX.html#strftime as in C, and treat the replace part as perl code with the e modifier, as per Embedding evaluations in Perl regex, and, ultimately, as documented at http://perldoc.perl.org/perlrequick.html#Search-and-replace, which makes it extra flexible as you can do further improvements to the date format (although on my system %v didn't work, and %e had extra space in one-digit days).
% cat 48941818.txt | perl -pe 'use POSIX qw(strftime); s#,\K([0-9]{4})-([01][0-9])-([0-3][0-9])T([0-9]{2}):[0-9]{2}:[0-9]{2}.[0-9]{3}Z(?=,|$)#POSIX::strftime(qw/%e-%b-%Y/,0,0,$4,$3,$2-1,$1-1900)#ge'
id,x1,x2,x3
AIR,Partner,20-Oct-2015,20-Oct-2015,12-Feb-2016
CMX,Partner,Tier,23-Mar-2017
WKA,Partner,Tier,22-May-2017
APP,Partner, 4-Oct-2017,Tier
BUN,27-Sep-2017,Partner,,27-Sep-2017
%