How can i set the delimiter as an argument in the match() function? - regex

Question

awk '{while(match($0,/("[^"]+",|[^,]*,|([^,]+$))/,a)){
$0=substr($0,RSTART+RLENGTH);b[++x]=a[0]}
print b[1] b[4];x=0}' file
I want to understand the match clause and want to know how can I make it dynamic so that it can take the delimiter as an argument instead of hard-coding it to comma.
I tried this, but it didnt work as i dont have the background for this function.
awk -v dl '{while(match($0,/("[^"]+"dl|[^,]*dl|([^,]+$))/,a)){
$0=substr($0,RSTART+RLENGTH);b[++x]=a[0]}
print b[1] b[4];x=0}' file
Input File data:
a,b,c,"d,e,f",
"a,b",c,d,"e,f",
p,q,r,"s,u",
Desired output (may be 4th field):
d,e,f
e,f
s,u
Desired output (may be 5th field, so it should generate the 3 rows with blank value):
Here , Delimiter can be anything comma, pipe and desired field number is also dynmaic.. thats why i wanted to pass the argument for field number and delimiter..
Field number argument is working fine but not the delimiter argument?
As suggested by Anubhava, i used that fpat which works really fine but it is not giving any rows when fetching the column 5th form the input file?

Answer 1

Using gnu-awk, you can define a FPAT variable that is a regular expression for matching fields.
awk -v FPAT='"[^"]*"|[^,]*' '{gsub(/"/, "", $4); print $4}' file
d,e,f
e,f
s,u
Running it from a shell script that takes delimiter as argument:
dl="${1?}"
awk -v FPAT='"[^"]*"|[^'"$dl"']*' '{gsub(/"/, "", $4); print $4}' "${2?}"
Then run it as:
bash p.sh ',' 'file'

Answer 2

That regex was strange, so I'll rewritten it. The regex:
/("[^"]+",|[^,]*,|([^,]+$))/
the "[^"]+" is parsed like - first " and last " are quotation marks and [^"]+ matches everything except quotes. So it's the same as:
"([^\"]+,|[^,]*,|([^,]+$))"
I guess that you want to match a field [^,]+ or a quoted field \"[^\"]+\" followed by a delimiter or end of line (,|$). So match that. And in matching groups match the insides of the fields, so match \"([^\"]+)\" or the unquoted field ([^,]+) and then use those matching groups
awk -v dl=, '{
x = 0;
while (match($0, "^(\"([^\"]+)\"|([^" dl "]+))(" dl "|$)", a)) {
$0 = substr($0, RSTART + RLENGTH);
b[++x] = a[2] a[3]; # funny, one of them will be empty
}
print b[4];
}' <<EOF
a,b,c,"d,e,f"
"a,b",c,d,"e,f"
p,q,r,"s,u"
EOF
d,e,f
e,f
s,u