Extract Fixed Columns from Pipelines with col1 to col9 and NF

Grab the third field of a pipeline with col3 instead of typing awk '{print $3}' for the hundredth time. The col1 through col9 commands select that numbered field, and NF selects the last one. They are small shell wrappers around awk, installed here by byobu 6.11-0ubuntu1.1, and this takes about ten minutes.

• You need: a shell and the byobu package.
• What it touches: nothing. The examples only read input and write to standard output. No edits, no services, no elevated privileges.

1. Check which helper is installed

Confirm the command is on your path and note the package version:

$ command -v col3
/usr/bin/col3
$ dpkg-query -W -f='${Package} ${Version}\n' byobu
byobu 6.11-0ubuntu1.1

Your package version can differ. The manual page installed with this version is dated 27 September 2010, and the helpers keep the simple interface described below.

2. Select a field from whitespace-separated input

With no argument, the helper splits on whitespace. The command name supplies the field number:

$ printf '%s\n' 'alpha 10 ready' 'beta 20 waiting' | col2
10
20
$ printf '%s\n' 'alpha 10 ready' 'beta 20 waiting' | col3
ready
waiting

Use col1 for the first field, col2 for the second, and so on through col9. It suits quick looks at the output of mount, ps or any command that emits stable whitespace-separated fields.

3. Select the last field with NF

NF is the awk term for the number of fields on the current record. Here it is a command that prints the last field of each input line:

$ printf '%s\n' 'alpha 10 ready' 'beta 20 waiting' | NF
ready
waiting

This is often safer than a high fixed number when the count of leading fields varies. It is still positional: if a line ends with a value you did not mean, NF prints that value.

4. Pass a delimiter for non-whitespace input

Give one optional argument to set the awk field separator. For comma-separated text, pass a comma:

$ printf '%s\n' 'alpha,10,ready' 'beta,20,waiting' | col3 ,
ready
waiting
$ printf '%s\n' 'alpha,10,ready' 'beta,20,waiting' | NF ,
ready
waiting

The separator goes to awk as its field-separator expression. It is not a CSV parser: it does not strip quoting or understand commas inside quoted fields. For simple delimited text that is enough. For real CSV, use a tool that understands the format.

The delimiter is consumed by the helper, and any later arguments go to awk as input files, so a file operand works neatly:

$ col3 : /path/to/status.txt
ready
waiting

Replace /path/to/status.txt with a readable file. No sudo is needed unless the file itself is restricted.

5. Check for missing fields

If a line has fewer fields than requested, the command prints a blank line and still exits successfully. That is just awk being asked for a field that is not there. It is not proof the source held an empty value:

$ printf '%s\n' 'one two' | col9 | od -An -t x1
 0a

The byte 0a is a newline with no field content. When an empty selection would be dangerous, guard the pipeline:

value=$(printf '%s\n' 'alpha 10 ready' | col3)
if [ -n "$value" ]; then
    printf 'selected: %s\n' "$value"
else
    printf '%s\n' 'expected field was missing' &2
    exit 1
fi

Use this check when the selected value controls a later command. For multi-line input, add a more explicit validation step if a blank line is not the only failure that matters.

6. Keep the field boundary visible

Before saving a command in a script, compare the helper with the equivalent awk expression. It makes the assumption reviewable:

$ printf '%s\n' 'alpha 10 ready' 'beta 20 waiting' | col3
ready
waiting
$ printf '%s\n' 'alpha 10 ready' 'beta 20 waiting' | awk '{print $3}'
ready
waiting

• Use col3 when the compact form reads better in a pipeline.
• Use explicit awk when you also need conditions, named variables, quoting-aware parsing or more than one selected field. The helper provides none of those.