You can do it by storing the numbers in an array that you output in an END block:
{
$1 = ""
count[++n] = gsub("o", "")
}
END {
$0 = "" # empty the current record
for (i = 1; i <= n; ++i)
$i = count[i] # create new fields for output
# output the current record (created above)
print
}
Or, you can output the numbers as you calculate them, separated by the field delimiter if needed, and then finish the line in the END block:
{
$1 = ""
# Print the number, preceded by OFS if this is not the first line.
# Does not output a newline (ORS).
printf "%s%s", (NR > 1 ? OFS : ""), gsub("o", "")
}
END {
# Finish the output (ORS is a newline by default).
printf "%s", ORS
}
Both of these variants allow the user to set OFS and ORS to values appropriate to them on the command line, e.g.
$ awk -v OFS=',' -f script.awk file
2,1
$ awk -v OFS='\t' -f script.awk file
2 1
$ awk -v OFS='-->' -v ORS='.\n' -f script.awk file
2-->1.
Note: I'm assuming that the output shown in the question is wrong as it's what you would get when not ignoring the first field (and ignoring the case of the o letter).