I have a shell script like this:
cat file | while read line
do
# run some commands using $line
done
Now I need to check if the line contains any non-whitespace character ([\n\t ]), and if not, skip it. How can I do this?
Add a comment
|
7 Answers
Since read reads whitespace-delimited fields by default, a line containing only whitespace should result in the empty string being assigned to the variable, so you should be able to skip empty lines with just:
[ -z "$line" ] && continue
5 Comments
Arkku
(More accurately, the delimiter used by
read is determined by the IFS variable, which defaults to whitespace. Just unset IFS to revert to using whitespace.)
pihentagy
even simpler: no need to quote line, if you use bash's [[ syntax:
[[ -z $line ]] && continueArkku
@pihentagy Umm, that has the same number of characters, and the square brackets
[] are harder to type than quotes on some international keyboards, and it becomes bash-specific. So maybe not simpler, but an alternative. =)
Xin Cheng
(for those newbie like me).
[ -z "$line" ] && continue itself is executable. This elegant line is equivalent to if [ -z "$line" ] ; then continue ; fi. By the way, don't forget to set IFS=" \t\n" at the beginning unless you don't want to skip 'tab'.Arkku
re. suggested edits that add things, please post your own answers instead – this answers the original question as it is, and works quite universally regardless of the shell, so I don't think the answer is improved by complicating it with more cases for specific shells and/or needs beyond the original question
try this
while read line;
do
if [ "$line" != "" ]; then
# Do something here
fi
done < $SOURCE_FILE
2 Comments
c0dem4gnetic
More information about the square bracket notation can be found at the man page of test
Timo
disadadvantage here: If the part in the if is long, you get a code difficult to read. Therefore the continue solution is always recommended.
bash:
if [[ ! $line =~ [^[:space:]] ]] ; then
continue
fi
And use done < file instead of cat file | while, unless you know why you'd use the latter.
answered Apr 5, 2010 at 11:38
Ignacio Vazquez-Abrams
804k160 gold badges1.4k silver badges1.4k bronze badges
cat i useless in this case if you are using while read loop. I am not sure if you meant you want to skip lines that is empty or if you want to skip lines that also contain at least a white space.
i=0
while read -r line
do
((i++)) # or $(echo $i+1|bc) with sh
case "$line" in
"") echo "blank line at line: $i ";;
*" "*) echo "line with blanks at $i";;
*[[:blank:]]*) echo "line with blanks at $i";;
esac
done <"file"
Comments
if ! grep -q '[^[:space:]]' ; then
continue
fi
answered Apr 5, 2010 at 11:47
Ignacio Vazquez-Abrams
804k160 gold badges1.4k silver badges1.4k bronze badges
Comments
blank=`tail -1 <file-location>`
if [ -z "$blank" ]
then
echo "end of the line is the blank line"
else
echo "their is something in last line"
fi
Comments
awk 'NF' file | while read line
do
# run some commands using $line
done
stole this answer to a similar question: Delete empty lines using sed
