Why can't I use cat in a shell script?

[kevdog]

I want to grep a particular comment or line from within a bunch of files. For example within a particular directory if I do the following:

# cat * | grep comment

This will find the line(s) containing the Regex comment contained within all the files in the directory.

I can't accomplish this however in a shell script. In fact the cat command doesn't seem to work will from within a shell script at all. I think I read somewhere the when used in shell scripts, every word is assigned a variable. Is there a work around?

[Phenax]

The way you are using cat is bad form (http://partmaps.org/era/unix/award.html#cat%5B/url).

This way works, is easier, and eliminates the redundancy and extra process of cat.


grep comment *
Don't want the filename prefix?


grep -h comment *
Or recursively


grep -R comment ./dir/

[LaRoza]

Works fine over here.


~$cat p.sh
#!/bin/bash

cat p.c | grep "main"
~$./p.sh
int main(void)
~$

[Phenax]

Works fine over here.


~$cat p.sh
#!/bin/bash

cat p.c | grep "main"
~$./p.sh
int main(void)
~$


Indeed it does work. Perhaps he is executing the script from another directory hoping to get the content of that directory?

[kevdog]

Here is an example:



#!/bin/bash

file=( `ls | tr '\n' ' '` )

echo -e "Size of array: ${#file[@]}\n"
#echo -e "Array: $file\n"

for filename in ${file[@]}
do
echo -e "Filename: $filename \n"
filecontent=( `cat $filename` ) <---Line doesnt work -- rather than each line being one element in the array, it seems each word on the line is an individual array element. Do I need like a field separator or something.

for line in ${filecontent[@]}
do
echo $line;
if [ $line = "comment" ]
then
echo $line
fi
done
done

echo -e "Done\n"



Also I want to grep an array for a value -- I know you can do it in perl, but how about in a shell script?

[rikxik]

Try this:



#!/bin/bash

file=( `ls | tr '\n' ' '` )

echo -e "Size of array: ${#file[@]}\n"
#echo -e "Array: $file\n"

for filename in ${file[@]}
do
echo -e "Filename: $filename \n"
OIFS=$IFS
IFS=$(echo "\006")
filecontent=( `cat $filename` )

for line in ${filecontent[@]}
do
echo $line;
if [ $line = "comment" ]
then
echo $line
fi
done
IFS=$OIFS
done

echo -e "Done\n"


Output:



Size of array: 2

Filename: a.txt

line 1 from a.txt
line 2 from a.txt
Filename: b.txt

line 1 from b.txt
line 2 from b.txt
Done



HTH

[ghostdog74]

Here is an example:



#!/bin/bash

file=( `ls | tr '\n' ' '` )

echo -e "Size of array: ${#file[@]}\n"
#echo -e "Array: $file\n"

for filename in ${file[@]}
do
echo -e "Filename: $filename \n"
filecontent=( `cat $filename` ) <---Line doesnt work -- rather than each line being one element in the array, it seems each word on the line is an individual array element. Do I need like a field separator or something.

for line in ${filecontent[@]}
do
echo $line;
if [ $line = "comment" ]
then
echo $line
fi
done
done

echo -e "Done\n"



Also I want to grep an array for a value -- I know you can do it in perl, but how about in a shell script?

let's leave arrays aside for a while. what you are doing above is long winded and uncommon way to parse a file. It can essentially be done using grep/awk as described by Phenax

[kevdog]

Yes I see that is a much more efficient way of doing things. So much to learn.