The definition of a field is that it is an unambiguous element so the concept of columns is unimportant. As long as the definition of each field is unique (ie, each field is separated by a : or , or / etc) then parsing the elements is trivial with many commands like sort, awk, cut, etc. To see this at work, run the cut command to find the user and the user's HOME directory (fields 1 and 6):
cut -f 1,6 -d: /etc/passwd
-d is the field delimiter. For awk, you specify the delimiter with -F and for sort the field delimiter is set with -t. awk is particularly useful as it predefines variables like NF (number of fields) so regardless of spacing, the last field can be extracted with $NF as in:
awk -F: '{print $1,$NF}' /etc/passwd
which prints the username and the shell.
Now your list may have fields defined with one delimiter such as a comma and within the field, data such as mm/dd/yy that uses / as a separator. That will be a problem because sort cannot understand nested fields. If you can show an example line, perhaps wew can suggest a way to sort it.
Bill Hassell, sysadmin
