grep last word of a given line .csv file

Question

I have a text file like this:

first, second
hello, bye
one, two
good, bad
day, night

I would like to have a script which, based on a variable, retrieves the last word of a given line using grep or awk.

for example: if num is 3, output should be two.

It's not necessary to ask for num's value.

Answer 1

Using awk:

$ awk -v num="$num" 'NR == num { print $NF }' data.in

Testing it:

$ num=3
$ awk -v num="$num" 'NR == num { print $NF }' data.in
two

The awk script reads the input file record by record (a record is by default a line). Once it hits the record corresponding to the num variable, it prints the last field of that record (a field is by default a whitespace-separated column).

The num variable inside the awk script is an awk variable that we initialize to the value of the shell variable num with -v num="$num" on the command line.

NR is the current record number, and NF is the number of fields in this record. $NF is the data of the last field.

If your file is strictly comma-separated, add -F ',' to the command line:

$ awk -v num="$num" -F ',' 'NR == num { print $NF }' data.in

With grep you can't select a specific line, but together with sed you can filter out the line you want and then get the last bit after the last comma:

$ sed -n "${num}p" data.in | grep -o '[^,]*$'

The sed bit will get the specified line while the grep bit will extract everything after the last comma on that line.

You may do it with sed only too:

$ sed -n "${num}s/^.*,\(.*\)$/\1/p" data.in

Here, the substitution is applied to only the line whose number is $num, and it replaces the whole line with the contents of the line after the last comma and outputs the result. All other output is inhibited with the -n command line switch.

Alternatively, use a substitution in sed to simply delete everything on the line to the last comma:

$ sed -n "${num}s/^.*,//p" data.in

Answer 2

In awk, NR is the line number, NF is the number of columns, so $NF is the last column. FS is the input record separator, ', ' for CSV without quoted commas should be enough.

awk -v FS=', ' 'NR==3{print $NF}'