A csv file with multiple records is delimited by |.


I want to check if field3 is blank or contains "space" characters only. If it is blank or space, the whole line should show up.

$ echo "1|2||4" | awk  -F'|' '$3 ~ /^[ \t]*$/   {print $0}'


$ echo "1|2|  |4" | awk  -F'|' '$3 ~ /^[ \t]*$/   {print $0}'

1|2|  |4

$ echo "1|2|  3|4" | awk  -F'|' '$3 ~ /^[ \t]*$/   {print $0}'

You can also use the cut command to pull out the third field and then test the value:

$ echo "field1|field2|field3|field4|field5" | cut -d '|' -f 3
  • More readable than awk for this trivial situation if y' ask me (which nobody did ;))
    – Tom Newton
    Jul 9 '09 at 7:08

My random attempt using grep would be:

grep -E '^[^|]*\|[^|]*\| *[^| ]+ *\|' file
  • The trouble with this approach is that it isn't general, so if you want to select on the 14th field it would become hateful.
    – James
    Jul 8 '09 at 23:27
  • 1
    @James: Then try something like: "grep -E '^([^|]*\|){13} *[^| ]+ *\|' file" adjusting the value "13" appropriately
    – JMusgrove
    Jul 13 '09 at 22:38

I'm not certain about unix, but in linux you would want to use the sed command.

sed 's/||/\n/g' will make it so that if there are any blank fields it will add a new line. not certain how to get it to only check the 3rd field. sed 's/| |/\n/g' should work for only spaces.

  • linux is a form of unix. Most things that are in linux are in other unixes. Jul 16 '09 at 12:40
  • was this comment necessary? Sed, Awk, Grep and the like are SIMILAR across *NIX and BSD systems, but not always the same. Jul 23 '09 at 19:53

Using Perl:

perl -F'\|' -lane 'print if $F[2] !~ /\S/' file

-a turns on autosplit mode, which splits the fields into array @F
-F'\|' sets the field delimiter to |
$F[2] is the 3rd field
!~ /\S/ tests for non-space characters (or empty)

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.