Question

I have some scripts which use awk to parse a CSV file. I have noticed that, if a cell is empty, awk simply moves to the next cell. This means, if I ask it to read column 4, but that cell is empty, it prints the data from column 5, e.g.:

echo "1@2@3@@5" | awk -F "@*" '{print $4}'

My expected result is that it will print nothing, because column 4 is empty.

• Why is awk skipping column 4?
• How can I get awk to not ignore empty columns?

Answer 1

The problem is not what you think. awk is not ignoring empty cells; it is parsing that line as 4 fields instead of 5.

[me@home]$ echo "1@2@3@@5" | awk -F "@*" '{print NF}'
4

That’s becuase you’re using @* as your field separator which allows one or more consecutive @ as your field separator (@, @@, @@@, … are all valid field separators).

Try using -F "@" instead.

[me@home]$ echo "1@2@3@@5" | awk -F "@" '{print NF}'
5
[me@home]$ echo "1@2@3@@5" | awk -F "@" '{print $4}'

[me@home]$ echo "1@2@3@@5" | awk -F "@" '{print $5}'
5