How does one properly call multiple cat commands from a single bash script?

Question

I am trying to collect text files from multiple subdirectories and concatenate them into a single file per subdirectory. There may be a much better way of doing this but my question is, how does one call multiple cat commands from a bash script? Here is an example:

#!/bin/bash
cat "./directory_1/"*.txt > "directory_1.txt"
cat "./directory_2/"*.txt > "directory_2.txt"

If I keep all the commands on the same line, seperated by semicolons, it works. For example:

#!/bin/bash
cat "./directory_1/"*.txt > "directory_1.txt";cat "./directory_2/"*.txt > "directory_2.txt"

But if I need to call say 20 directories, the script becomes very hard to read.

I expected the first example to run each cat command, however, the following error was thrown:

line 2: $'\r': command not found
: No such file or directorydirectory2.txt

It looks like I have a problem with the newline character. I am trying to run this command on the Ubuntu shell for Windows 10.

Answer 1

The following loop would process every directory under the current directory in the way that you describe:

for dir in */; do
    cat "$dir"/*.txt >"${dir%/}.txt"
done

The ${dir%/} thing removes the trailing / from $dir which is there from the glob match in the loop header.

The error that you get is likely due to writing the script on a Windows machine (with a Windows editor). Windows (and DOS) text files has a carriage return character before the linefeed character at the end of every line, while Unix text files only has the linefeed character. You can convert the script to a Unix text file using the dos2unix utility.

Answer 2

The error message is spot on in this case. The script was being saved with Windows style line endings "CR/LF" instead of Unix style "LF." Updating the line endings of the bash file fixed the problem.