Change FS definition past the header of a CSV file - regex

Question

I got this kind of CSV file:
name,x-extension,value,extra
"Roger","9890","",""
"Nicole","9811","president, ceo",""
...
Now, I want the find the maximum size of each field in the file. So I used this awk script:
Updated script:
NR==1 {
gsub(/\r/,"",$0) #remove linefeed from last field name
for (n = 1; n <= NF; n++) {
colname[n]=$n;
maxlen[n]=-1;
}
nbrField = NF; # will get bump +2 by the new FS
FS="\",\"|^\"|\"$";
}
NR>1 {
for (n = 2; n <= nbrField+1; n++) {
if (length($n)>maxlen[n-1]) {
maxlen[n-1]=length($n);
}
}
}
END {
for(i = 1; i <= nbrField; i++) {
printf "%s : %s\n", colname[i], maxlen[i]
}
}
The problem a got is I need to change the field separator AFTER reading the first line because as you can see, the header don't use double quote for field delimiter and there is coma INSIDE some field.
I tried to play with this -F option on my awk command line but I can't find the right combination of regex to do the trick..
> awk -F'", "|^"|"$' -f myprog mydata ==>(don't work)
Help! :-)

Answer 1

Change FS in the block that processes the first line:
NR==1 {
for(n = 1; n <= NF; n++) {
colname[n]=$n
}
FS="\",\"|^\"|\"$"
}

Answer 2

I prefer to use a real CSV parser for CSV data. For example, Perl:
perl -MText::CSV -MList::Util=max -nE '
BEGIN {$csv = Text::CSV->new({binary=>1})}
$csv->parse($_);
#row = $csv->fields();
if ($. == 1) {
#h = #row; $n = $#row;
}
else {
$max{$h[$_]} = max($max{$h[$_]}, length $row[$_]) for (0..$n)
}
END {
while (($k,$v) = each %max) {say join ":", $k, $v}
}
' << DATA
name,x-extension,value,extra
"Roger","9890","",""
"Nicole","9811","president, ceo",""
DATA
value:14
name:6
extra:0
x-extension:4