dupline

# duplicates lines based on splitting on [,;]. Expects a filename and case insensitive column header as arguments

infile="${1}"
fileroot=$(echo "${infile}" | sed 's/\.....\?$//')
outfile="${fileroot}_fixed.tsv"

if [[ -z $2 ]];then echo "You must indicate a split column";exit;fi

split_column="${2//[^A-Za-z0-9]/}"
rm -f "${outfile}"

read -r -d '' awkscript << "ENDOFAWK"
#!/usr/bin/awk -f 

BEGIN {
	gsub(/[^a-z0-9]/, "", COL)
}
{
	if (NR == 1) {
		for(i=1;i<=NF;i++) { 
			compareheader=tolower($i)
			gsub(/[^a-z0-9]/, "", compareheader)
			if (compareheader == COL ) { dupcolumn = i }
		}
	print $0 > OUTFILE
	} else {
		recordsprocessed++

		numdups = patsplit($dupcolumn,col_array,/[^,;]+/)
		for (i=1;i<=numdups;i++) {
			recordscreated++
			for(j=1;j<=NF;j++) { 
				if (j == NF) {delim = RS} else {delim = OFS}
				if (j != dupcolumn ) {
					printf($j""delim) >> OUTFILE
				} else {
				#	printf("%s%s",$j,j==NF?RS:OFS) >> OUTFILE 
					printf(col_array[i]""delim) >> OUTFILE 
				}
			}
		}
	printf "Records processed: %d :created: %d \\r", recordsprocessed, recordscreated
	}
}
END { print "" } 

ENDOFAWK


echo -e "${awkscript}" > tmp_colsplit
chmod 700 tmp_colsplit

awk -v COL="${split_column}" -v OUTFILE="${outfile}" -v FS="\t" -v OFS="\t" -f tmp_colsplit "${infile}" 

rm tmp_colsplit

#echo "File has been rebuilt with multivalues in ${split_column} as individual rows and sent to ${outfile}"