Bash looping directories

bash

My simple task is to create within the workdir up to max separate dirs

copy within it some files from the template dir (so every new dir is replicas and the difference only within it names defined from i)

create new txt file within each of the new dir with some text:

home=/projects/clouddyn/md_bench/for_server/MD_bench
template=$home/template

cd $home
min=1
max=3

for i in $(eval echo "{$min..$max}")
do
    mkdir "sim_$i"
    sim=$(basename "sim_$i")
    pushd $sim
    cp $template/*.* $sim
    prinf "This is $i dir" > ./$sim/file_$i.txt
    popd
done

Unfortunately the new dirs are created but files didn't copy

Thanks for the help!

Best Answer

There are several things in your script which can make it go wrong:

#!/bin/bash
home=/projects/clouddyn/md_bench/for_server/MD_bench

# home is a bad choice, because $HOME has a special meaning and it's
# confusing to have both $HOME and $home. But it doesn't make the script fail

template=$home/template

cd $home
min=1
max=3

# for i in $(eval echo "{$min..$max}")
# this can be replaced by the easier to read
for i in {$min..$max}
do
    mkdir "sim_$i"

    # sim=$(basename "sim_$i")
    # basename strips the path part, but "sim_$i" doesn't even have a path
    # so basically this just assigned "sim_$i" to $sim and can be replaced by
    sim="sim_$i"
    # Ideally this assignment would be the first statement after 'do' so 
    # it could be used already in 'mkdir'

    # pushd $sim
    # cp $template/*.* $sim
    # Two problems here:
    # - 'pushd' already changes into $sim, so when 'cp' is called there is
    #   no $sim directory to copy into (as you are already in it)
    # - '$template/*.*' only matches files with a dot in the name, which may
    #   or may not be what you want
    # As there is no need to 'cd' into a directory to copy files into it I
    # would just use
    cp $template/* $sim/

    # prinf "This is $i dir" > ./$sim/file_$i.txt
    # You probably meant 'printf' here. Also this line has the same problem
    # as the 'cp' above (you are already in $sim)
    printf "This is $i dir" > $sim/file_$i.txt

    # popd
    # Not needed any longer
done

Related Solutions

Bash: Searching with ls

To find all files which contain a certain string, you can use just grep on its own:

grep -R /path/to/directory "myword"

In the example above, "myword" is what you're looking for, and /path/to/directory is the directory you want to search through. -R tells grep to run a recursive search (i.e. to start at the path you specified, and go into each subdirectory in there).

EDIT

I've re-read your question and I think you're looking for how to find files that have a certain filename. For this use:

find /path/to/directory -name "*glob*" -print

where /path/to/directory is where you want to start your search (and find will also look in the subdirectories) and "*glob*" is a full or a partial file name that you want to look for. Please note, you can use wildcards (. and * for any single and any multiple characters match, respectively) in the glob. As a variation to this, you can use, as proposed by others:

find /path/to/directory -name "*glob*" -exec ls -la {} \;

in the above, what you put after the option -exec is any UNIX command with paramteres (in this case ls -la). The {} will be substituted by the file name find finds, and the \; is a termination string.

EDIT 2

Please note, the commands above give you similar results to Spotlight, but they don't use Spotlight at all.

If you'd like to use Spotlight from command line, you can use mdfind command, which is a command-line interface for Spotlight. Typing mdfind on its own will give you a quick help for using it with a few examples. man mdfind will give you a more detailed manual for it.

Essentially mdfind takes a search string (and optional modifiers) as it's argument and it's the same as you would type in Spotlight's text field when using it in a GUI.

Remove special characters and replace space with _ from files and directories Recursively in Bash

Untested

find . -depth -print0 |
while IFS= read -d '' -r file; do
    dir=$(dirname "$file")
    base=$(basename "$file") 
    base=${base//[[:space:]]/_}         # change whitespace to underscores
    base=${base//[^[:alnum:]_.]/}       # delete any chars not alphanumeric _ .
    newname="$dir/$base"
    if [[ -e $newname ]]; then
        echo "'$newname' exists, not renaming '$file'"
    else
        echo mv "$file" "$newname"
    fi
done

If it looks OK, remove the "echo"

The key here is the "-depth" option to find. It will emit "./sub dir/file name" before "./sub dir". Otherwise if "./sub dir" is renamed to "./subdir", then you'll get a 'file not found' error when the while loops gets to "./sub dir/file name". Also, it means you don't need different logic for directories.

The "-print0" option to find and the busy-ness around the read command are to ensure filenames with spaces are handled properly.

Best Answer

Related Solutions

Bash: Searching with ls

Remove special characters and replace space with _ from files and directories Recursively in Bash

Related Question