If you save the following awk script in a file named Serials.awk:

Code:

BEGIN {	FS = OFS = ","
}
FNR == NR {
	n[$1,++nc[$1]] = $2
	next
}
FILENAME == "Serials.csv" {
	nextfile
}
FNR == 1 {
	if(of != "")
		close(of)
	of = FILENAME ".out"
}
{	if(!($1 in nc))
		print $1, "absent", $2 > of
	else {	print $1, n[$1, 1], $2 > of
		if(FNR == 3)
			for(i = 2; i <= nc[$1]; i++)
				print $1, n[$1, i], $2 > of
	}
}

Then the awk command:

Code:

awk -f Serials.awk Serials.csv *.csv

might be all that you need. If it gives you an argument list too long error, try the following instead:

Code:

ls | grep '.csv$' | xargs awk -f Serials.awk Serials.csv

If you want to try this on a Solaris/SunOS system, change the awk in the above commands to /usr/xpg4/bin/awk.

If the awk utility on your system does not support the nextfile function, change the line in Serials.awk that currently says:

Code:

	nextfile

to:

Code:

	next

The above script produces the output you requested, except the order of the a/b lines is different in some of the *.out files, and the line:

Code:

namec3,s8,,a/b

in File3.csv.out comes out as:

Code:

namec3,s8,a/b

which seems to be what you requested.