How can I prevent the slash from crashing my program. My first thought is to remove the slash but if there is a better method, please let me know. I don't know if the user will pass in "dir" or "dir/". This is in sh.
$1="directory/"
for i in "$1"/*
do
some code
done
You can remove it if present with a parameter expansion:
for i in "${1%/}"/*
which expands to $1 with the slash removed from the end.
I think you have a problem like #Benjamin said: Do not assign to $1.
Let's check the input and assign to dir.
#
if [ $# -eq 0 ]; then
echo "Usage: $0 directory [other parameters]"
exit 1
fi
if [ -d "$1"]; then
dir="${1%/}"
else
echo "$1 is not a valid directory"
exit 1
fi
for file in "$dir"/*
do
some code with "${file}"
done
Related
I want to remove the extension of specific files with a given extension.
So for instance, in a directory foobar, we have foo.txt, bar.txt foobar.jpg.
Additionally, the extension that I've put in to be removed is txt
After calling the program, my output should be foo bar foobar.jpg
Here is my code so far:
#!/bin/bash
echo "Enter an extension"
read extension
echo "Enter a directory"
read directory
for file in "$directory"/*; do //
if [[ $file == *.txt ]]
then
echo "${file%.*}"
else
echo "$file"
fi
done
However when I run this on a given directory, nothing shows up.
I'm assuming that there is a problem with how I referred to the directory ( in the line where I placed a //) and I've tried to research on how to solve it but to no avail.
What am I doing wrong?
If files do exist in a valid directory you've entered then they should show up — with one exception. If you are using ~/ (shorthand home directory) then it will be treated as plain text in your for loop. The read variable should be substituted into another variable so the for loop can treat it as a directory (absolute paths should work normally as well).
#!/bin/bash
echo "Enter an extension"
read -r extension
echo "Enter a directory"
read -r directory
dir="${directory/#\~/$HOME}"
for file in "$dir"/*; do
if [[ $file == *."$extension" ]]
then
echo "${file%.*}"
else
echo "$file"
fi
done
You can simplify your for-loop:
for file in "$directory"/*; do
echo "${f%.$extension}";
done
The % instructions removes only matching characters. If nothing matches, the original string (here f) is returned.
When you write bash scripts it's more common to pass arguments to your script via command line arguments rather than by reading it from standard input via read program.
Passing arguments via command line:
#!/bin/bash
# $# - a bash variable which holds a number of arguments passed
# to script via command line arguments
# $0 holds the name of the script
if [[ $# -ne 2 ]]; then # checks if exactly 2 arguments were passed to script
echo "Usage: $0 EXTENSION DIRECTORY"
exit -1;
fi
echo $1; # first argument passed to script
echo $2; # second arugment passed to script
This approach is more efficient because a subprocess is spawn for read command to run and there is no subprocess spawn for reading command line arguments.
There is no need to manually loop through directory, you can use find command to find all files with given extension within given directory.
find /path/to/my/dir -name '*.txt'
find $DIRECTORY -name "*.$EXTENSION"
# note that single quotes in this context would prevent $EXTENSION
# variable to be resolved, so double quotes are used " "
# find searches for files inside $DIRECTORY and searches for files
# matching pattern '*.$EXTENSION'
Note that to avoid bash filename expansion sometimes it is required to wrap actual pattern in single quotes ' ' or double quotes " ". See Bash Filename Expansion
So now your script can look like this:
#!/bin/bash
if [[ $# -ne 2 ]]; then
echo "Usage: $0 EXTENSION DIRECTORY"
exit -1;
fi
$EXTENSION = $1 # for better readability
$DIRECTORY = $2
for file in `find $DIRECTORY -name "*.$EXTENSION"`; do
mv $file ${file%.$EXTENSION}
done
Construct ${file%.$EXTENSION} is called Shell Parameter Expansion it searches for occurrence of .$EXTENSION inside file variable and deletes it.
Notice that in the script it is easy to pass extension as directory and vice versa.
We can check if second argument is in fact directory, we can use following construction:
if ! [[ -d $DIRECTORY ]]; then
echo $DIRECTORY is not a dir
exit -1
fi
This way we can exit from the script earlier with more readable error.
To sum up entire script could look like this:
#!/bin/bash
if [[ $# -ne 2 ]]; then
echo "Usage: $0 EXTENSION DIRECTORY"
exit -1;
fi
EXTENSION=$1 # for better readability
DIRECTORY=$2
if ! [[ -d $DIRECTORY ]]; then
echo $DIRECTORY is not a directory.
exit -1
fi
for file in `find $DIRECTORY -name "*.$EXTENSION"`; do
mv $file ${file%.$EXTENSION}
done
Example usage:
$ ./my-script.sh txt /path/to/directory/with/files
This question already has an answer here:
Shell spacing in square brackets [duplicate]
(1 answer)
Closed 4 years ago.
$1 is file / folder that want to compressed
Output filename is the same name, plus current date and ext
if output name exist, then just give warning
Example:
./cmp.sh /home/user
It will be /home/user to /home/user_2018-03-11.tar.bz2
i already have lead, but i'm stuck
#!/bin/bash
if ["$1" == ""]; then
echo "Help : To compress file use argument with directory"
exit 0
fi
if [[ -f "$1" || -d "$1" ]]; then
tar -cvjSf $1"_"$(date '+%d-%m-%y').tar.bz2 $1
fi
but the output is _22-04-2018.tar.bz2
I see that you're using quotes to avoid the problem the underscore getting used as part of the variable name. So while $1 is a positional paramater, $1_ is a variable that you have not set in your script. You can avoid this issue by using curly braces, like ${1}. Anything inside the braces is part of the variable name, so ${1}_ works. This notation would be preferable to $1"_" which leaves a user-provided variable outside of quotes. (Of course, "$1"_ would do the job as wel.)
Also, it's probably safer to set the filename in a variable, then use that for all your needs:
#!/bin/bash
if [ -z "$1" ]; then
echo "Help : To compress file use argument with directory"
exit 0
fi
filename="${1}_$(date '+%F').tar.bz2"
if [ -e "$filename" ]; then
echo "WARNING: file exists: $filename" >&2
else
tar -cvjSf "$filename" "$#"
fi
Changes:
you need spaces around your square brackets in an if condition,
while you can test for equivalence to a null string, -z is cleaner, though you could also test for [ $# -eq 0 ], counting the parameters provided,
using $filename makes sure that your test and your tar will always use the same name, even if the script runs over midnight, and is way more readable,
variables should always be quoted.
Also, are you sure about the -S option for tar? On my system, that option extracts sparse files, and is only useful in conjunction with -x.
ALSO, I should note that as I've rewritten it, there's nothing in this script which is specific to bash, and it should be portable to POSIX shells as well (ash/dash/etc). Bash is great, but it's not universal, and if through your learning journey you can learn both, it will give you useful skills across multiple operating systems and environments.
Use -z switch to check if blank
#!/bin/bash
if [[ -z "$1" ]]; then
echo "Help : To compress file use argument with directory"
exit 0
fi
if [[ -f "$1" || -d "$1" ]]; then
tar -cvjSf $1"_"$(date '+%d-%m-%y').tar.bz2 $1
fi
My requirement is simple
I have a file with a list of paths provided in a line.
Eg:
/etc/home
/var/www/
I need a command to validate that there are no space characters in between the path like: /var/ www/
How can I achieve this validation?
This should help. It uses the Internal Field Separator.
#!/bin/sh
if [ ! $1 ]; then
echo Filename parameter expected
exit
fi
export IFS=$'\n'
for i in `cat $1`
do
if [ `echo $i | grep ' ' ` ]; then
echo Spaces found in [$i]
else
echo No spaces found in [$i]
fi
done
How to write a bash shell script called 'abc' which takes one argument, the name of a directory, and adds the extension ".xyz" to all visible files in the directory that don't already have it
I have mostly written the code which changes the filenames inside the current directory but I can't get the script to accept an argument (directory name) and change the filenames of that directory
#!/bin/bash
case $# in
0) echo "No directory name provided" >&2 ; exit 1;;
1) cd "${1}" || exit $?;;
*) echo "Too many parameters provided" >&2 ; exit 1;;
esac
for filename in *
do
echo $filename | grep "\.xyz$"
if [ "$?" -ne "0" ]
then mv "$filename" "$filename.old"
fi
done
additional instructions include;
Within 'abc', use a "for" control structure to loop through all the non-hidden filenames
in the directory name in $1. Also, use command substitution
with "ls $1" instead of an ambiguous filename, or you'll descend into subdirectories.
EDIT: The top part of the question has been answered below, however the second part requires me to modify my own code according to the following instructions:
Modify the command substitution that's being used to create the loop values that will be placed into the "filename" variable. Instead of just an "ls $1", pipe the output into a "grep". The "grep" will search for all filenames that DO NOT end in ".xyz". This can easily be done with the "grep -v" option. With this approach, you can get rid of the "echo ... | grep ..." and the "if" control structure inside the loop, and simply do the rename.
How would I go about achieving this because according to my understanding, the answer below is already only searching through filenames without the .xyz extension however it is not being accepted.
Your description is a little unclear in places, so I've gone with the most obvious:
#!/bin/bash
# validate input parameters
case $# in
0) echo "No directory name provided" >&2 ; exit 1;;
1) cd "${1}" || exit $?;;
*) echo "Too many parameters provided" >&2 ; exit 1;;
esac
shopt -s extglob # Enables extended globbing
# Search for files that do not end with .xyz and rename them (needs extended globbing for 'not match')
for filename in !(*.xyz)
do
test -f && mv "${filename}" "${filename}.xyz"
done
The answer to the second part is this:
#!/bin/bash
for file in $(ls -1 "$1" | grep -v '\.old$'); do
mv "$file" "$file.old"
done
I got it from somewhere
#!/bin/bash
hello()
{
SRC=$1
DEST=$2
for IP in `cat /opt/ankit/configs/machine.configs` ; do
echo $SRC | grep '*' > /dev/null
if test `echo $?` -eq 0 ; then
for STAR in $SRC ; do
echo -en "$IP"
echo -en "\n\t ARG1=$STAR ARG2=$2\n\n"
done
else
echo -en "$IP"
echo -en "\n\t ARG1=$SRC ARG2=$DEST\n\n"
fi
done
}
hello $1 $2
The above is the shell script which I provide source (SRC) & desitnation (DEST) path. It worked fine when I did not put in a SRC path with wild card ''. When I run this shell script and give ''.pdf or '*'as follows:
root#ankit1:~/as_prac# ./test.sh /home/dev/Examples/*.pdf /ankit_test/as
I get the following output:
192.168.1.6
ARG1=/home/dev/Examples/case_Contact.pdf ARG2=/home/dev/Examples/case_howard_county_library.pdf
The DEST is /ankit_test/as but DEST also get manupulated due to '*'. The expected answer is
ARG1=/home/dev/Examples/case_Contact.pdf ARG2=/ankit_test/as
So, if you understand what I am trying to do, please help me out to solve this BUG.
I'll be grateful to you.
Thanks in advance!!!
I need to know exactly how I use '*.pdf' in my program one by one without disturbing DEST.
Your script needs more work.
Even after escaping the wildcard, you won't get your expected answer. You will get:
ARG1=/home/dev/Examples/*.pdf ARG2=/ankit__test/as
Try the following instead:
for IP in `cat /opt/ankit/configs/machine.configs`
do
for i in $SRC
do
echo -en "$IP"
echo -en "\n\t ARG1=$i ARG2=$DEST\n\n"
done
done
Run it like this:
root#ankit1:~/as_prac# ./test.sh "/home/dev/Examples/*.pdf" /ankit__test/as
The shell will expand wildcards unless you escape them, so for example if you have
$ ls
one.pdf two.pdf three.pdf
and run your script as
./test.sh *.pdf /ankit__test/as
it will be the same as
./test.sh one.pdf two.pdf three.pdf /ankit__test/as
which is not what you expect. Doing
./test.sh \*.pdf /ankit__test/as
should work.
If you can, change the order of the parameters passed to your shell script as follows:
./test.sh /ankit_test/as /home/dev/Examples/*.pdf
That would make your life a lot easier since the variable part moves to the end of the line. Then, the following script will do what you want:
#!/bin/bash
hello()
{
SRC=$1
DEST=$2
for IP in `cat /opt/ankit/configs/machine.configs` ; do
echo -en "$IP"
echo -en "\n\t ARG1=$SRC ARG2=$DEST\n\n"
done
}
arg2=$1
shift
while [[ "$1" != "" ]] ; do
hello $1 $arg2
shift
done
You are also missing a final "done" to close your outer for loop.
OK, this appears to do what you want:
#!/bin/bash
hello() {
SRC=$1
DEST=$2
while read IP ; do
for FILE in $SRC; do
echo -e "$IP"
echo -e "\tARG1=$FILE ARG2=$DEST\n"
done
done < /tmp/machine.configs
}
hello "$1" $2
You still need to escape any wildcard characters when you invoke the script
The double quotes are necessary when you invoke the hello function, otherwise the mere fact of evaluating $1 causes the wildcard to be expanded, but we don't want that to happen until $SRC is assigned in the function
Here's what I came up with:
#!/bin/bash
hello()
{
# DEST will contain the last argument
eval DEST=\$$#
while [ $1 != $DEST ]; do
SRC=$1
for IP in `cat /opt/ankit/configs/machine.configs`; do
echo -en "$IP"
echo -en "\n\t ARG1=$SRC ARG2=$DEST\n\n"
done
shift || break
done
}
hello $*
Instead of passing only two parameters to the hello() function, we'll pass in all the arguments that the script got.
Inside the hello() function, we first assign the final argument to the DEST var. Then we loop through all of the arguments, assigning each one to SRC, and run whatever commands we want using the SRC and DEST arguments. Note that you may want to put quotation marks around $SRC and $DEST in case they contain spaces. We stop looping when SRC is the same as DEST because that means we've hit the final argument (the destination).
For multiple input files using a wildcard such as *.txt, I found this to work perfectly, no escaping required. It should work just like a native bash app like "ls" or "rm." This was not documented just about anywhere so since I spent a better part of 3 days trying to figure it out I decided I should post it for future readers.
Directory contains the following files (output of ls)
file1.txt file2.txt file3.txt
Run script like
$ ./script.sh *.txt
Or even like
$ ./script.sh file{1..3}.txt
The script
#!/bin/bash
# store default IFS, we need to temporarily change this
sfi=$IFS
#set IFS to $'\n\' - new line
IFS=$'\n'
if [[ -z $# ]]
then
echo "Error: Missing required argument"
echo
exit 1
fi
# Put the file glob into an array
file=("$#")
# Now loop through them
for (( i=0 ; i < ${#file[*]} ; i++ ));
do
if [ -w ${file[$i]} ]; then
echo ${file[$i]} " writable"
else
echo ${file[$i]} " NOT writable"
fi
done
# Reset IFS to its default value
IFS=$sfi
The output
file1.txt writable
file2.txt writable
file3.txt writable
The key was switching the IFS (Internal Field Separator) temporarily. You have to be sure to store this before switching and then switch it back when you are done with it as demonstrated above.
Now you have a list of expanded files (with spaces escaped) in the file[] array which you can then loop through. I like this solution the best, easiest to program for and easiest for the users.
There's no need to spawn a shell to look at the $? variable, you can evaluate it directly.
It should just be:
if [ $? -eq 0 ]; then
You're running
./test.sh /home/dev/Examples/*.pdf /ankit_test/as
and your interactive shell is expanding the wildcard before the script gets it. You just need to quote the first argument when you launch it, as in
./test.sh "/home/dev/Examples/*.pdf" /ankit_test/as
and then, in your script, quote "$SRC" anywhere where you literally want the things with wildcards (ie, when you do echo $SRC, instead use echo "$SRC") and leave it unquoted when you want the wildcards expanded. Basically, always put quotes around things which might contain shell metacharacters unless you want the metacharacters interpreted. :)