Reputation: 87
I have the following situation:
I have a text file I'm trying to loop so I can know if each line has a match with ".mp3" in this case which is this one:
12 Stones.mp3
randomfile.txt
Aclarion.mp3
ransomwebpage.html
Agents Of The Sun.mp3
randomvideo.mp4
So, I've written the following script to process it:
while read line || [ -n "$line" ]
do
varline=$(awk '/.mp3/{print "yes";next}{print "no"}')
echo $varline
if [ "$varline" == "yes" ]; then
some-command
else
some-command
fi
done < file.txt
The expected output would be:
yes
no
yes
no
yes
no
Instead, it seems misses the first line and I get the following:
no
yes
no
yes
no
Upvotes: 2
Views: 1527
Reputation: 5252
Have you forgot something? Your awk
has no explicit input, change to this instead:
while IFS= read -r read line || [ -n "$line" ]
do
varline=$(echo "$line" | awk '/.mp3/{print "yes";next}{print "no"}')
echo $varline
if [ "$varline" == "yes" ]; then
some-command
else
some-other-command
fi
done < file.txt
In this case, you might need to change to /\.mp3$/
or /\.mp3[[:space:]]*$/
for precise matching.
Because .
will match any character, so for example /.mp3/
will match Exmp3but.mp4
too.
Update: changed while read line
to while IFS= read -r read line
, to keep each line's content intact when assigning to the variable.
And the awk
part can be improved to:
awk '{print $0~/\.mp3$/ ? "yes":"no"}'
So with awk
only, you can do it like this:
awk '{print $0~/\.mp3$/ ? "yes":"no"}' file.txt
Or if your purpose is just the commands in the if structure, you can just do this:
awk '/\.mp3$/{system("some-command");next}{system("some-other-command");}' file.txt
or this:
awk '{system($0~/\.mp3$/ ? "some-command" : "some-other-command")}' file.txt
Upvotes: 2
Reputation: 189689
You really don't need Awk for a simple pattern match if that's all you used it for.
while IFS= read -r line; do
case $line in
*.mp3) some-command;,
*) some-other-command;;
esac
done <file.txt
If you are using Awk anyway for other reasons, looping the lines in a shell loop is inefficient and very often an antipattern. This doesn't really fix that, but at least avoids executing a new Awk instance on every iteration:
awk '{ print ($0 ~ /\.mp3$/) ? "yes" : no" }' file.txt |
while IFS= read -r whether; do
case $whether in
'yes') some-command ;;
'no') some-other-command;;
esac
done
If you need the contents of "$line"
too, printing that from Awk as well and reading two distinct variables is a trivial change.
I simplified the read
expression on the assumption that you can make sure your input file is well-formed separately. If you can't do that, you need to put back the more-complex guard against a missing newline on the last line in the file.
Upvotes: 4
Reputation: 3183
Use awk
$ awk '{if ($0 ~ /mp3/) {print "yes"} else {print "no"}}' file.txt
yes
no
yes
no
yes
no
Or more concise:
$ awk '/mp3/{print "yes";next}{print "no"}' file.txt
$ awk '{print (/mp3/ ? "yes" : "no")}' file.txt
Upvotes: 2