I am writing a menu based bash script, one of the menu options is to send an email with a text file attachment. I am having trouble with checking if my file is a text file. Here is what I have:

until [ $fileExists -eq 9 ]
  echo "Please enter the name of the file you want to attach: "
  read attachment
  isFile=$(file $attachment | cut -d\ -f2)
  if [[ $isFile = "ASCII" ]]
      echo "$attachment is not a text file, please use a different file"

I keep getting the error cut: delimiter must be a single character.

closed as off-topic by Michael Homer, G-Man, roaima, Stephen Kitt, PersianGulf Jun 11 '15 at 15:40

This question appears to be off-topic. The users who voted to close gave this specific reason:

  • "Questions describing a problem that can't be reproduced and seemingly went away on its own (or went away when a typo was fixed) are off-topic as they are unlikely to help future readers." – Michael Homer, G-Man, roaima, Stephen Kitt, PersianGulf
If this question can be reworded to fit the rules in the help center, please edit the question.

  • 3
    Put an extra space after -d\ . – Michael Homer Jun 11 '15 at 9:34
  • 1
    Depending on the file version you have available you should consider using some options like --brief (which doesn't output the filename so you will have less of a problem with filenames that contain spaces) or --mime which returns the MIME type (e.g. text/plain) instead of a textual description of the file type. – Dubu Jun 11 '15 at 10:00
  • 2
    Just a note on the off-topic closure - This question would still help a lot of future readers like me. I was looking for an if statement to check if a file contained text, and this one helped me perfectly. – thepiercingarrow Mar 16 '16 at 23:58

The problem occurs in cut -d\ -f2. Change it to cut -d\ -f2.

To cut, the arguments look like this:

# bash: args(){ for i; do printf '%q \\\n' "$i"; done; }
# args cut -d\ -f2
cut \
-d\ -f2 \

And here is the problem. \ escaped the space to a space literal instead of a delimiter between arguments in your shell, and you didn't add an extra space so the whole -d\ -f2 part appears as one argument. You should add one extra space so -d\ and -f2 appear as two arguments.

To avoid confusion, many people use quotes like -d' ' instead.

P.S.: Instead of using file and making everything ASCII, I'd rather use

if file "$attachment2" | grep -q text$; then
    # is text
    # file doesn't think it's text
  • ah thank you! It was my spaces again, but it's working now! – Powea Jun 11 '15 at 18:39
  • @Powea Added an alternative. – Arthur2e5 Jun 12 '15 at 10:35
  1. From the fact that it says file $attachment rather than file "$attachment", I guess your script cannot handle filenames that contain spaces.  But, be advised that filenames can contain spaces, and well-written scripts can handle them.  Note, then:

    $ file "foo bar"
    foo bar:  ASCII text
    $ file "foo bar" | cut -d' ' -f2

    One popular and highly recommended approach is to null-terminate the filenames:

    $ file -0 "foo bar" | cut -d $'\0' -f2
    :  ASCII text
  2. The file command makes educated guesses about what type of file a file is.  Guesses, naturally, are sometime wrong.  For example, file will sometimes look at an ordinary text file and guess that it is a shell script, C program, or something else.  So you don't want to check whether the output from file is ASCII text, you want to see whether it says that the file is a text file.  If you look at the man page for file, you will see that it more-or-less promises to include the word text in its output if the file is a text file, but this might be in a context like shell commands text.  So, it may be better to check whether the output from file contains the word text:

    isFile=$(file -0 "$attachment" | cut -d $'\0' -f2)
    case "$isFile" in
          echo "$attachment is a text file"
          echo "$attachment is not a text file, please use a different file"
  • 3
    You cannot rely on substrings in the output of file as for many formats file extracts and displays strings from the file (look for %s in the magic sources) which may include text – Stéphane Chazelas Jun 12 '15 at 22:36
  • @StéphaneChazelas: Perhaps you should take the wisdom that you have sprinkled over this page (-b, < filename, --mime) and add it to your answer, rather than scribbling it on leaves and letting them blow into everybody else’s yards.  :-)  ⁠ – G-Man Jun 16 '15 at 23:52
case $(file -b --mime-type - < "$attachment") in
     printf '%s\n' "$attachment is probably text according to file"
     case $(file -b --mime-encoding - < "$attachment") in
       (us-ascii) echo "and probably in ASCII encoding"

I would circumvent the escaping and do:

... | cut -d' ' -f2 

that way it is clear that you need a space between the delimiter character (specified by the three letters sequence ' ') and the following option. With -d\ -f2 it is easy to miss you should have done -d\ -f2.


Another option is to not use cut and to match a regex against the full output of file:

isFile=$(file $attachment)
if [[ "$var" =~ ^[^:]*:\ ASCII ]]
  • 1
    The elephant in the room is that the output from file is of the form “(filename) (colon) (whitespace) (filetype)” — i.e., it includes the filename.  Therefore, if there is a binary file called VACASCIING.JPG, the output from file will be VACASCIING.JPG: JPEG image, and your code will call it a text file because the filename matches ASCII.  Note that the second part of my answer, which is functionally comparable to yours,   … (Cont’d) – G-Man Jun 12 '15 at 18:57
  • 1
    (Cont’d) …  would not have needed the cut command but for this.  (The OP is using cut to extract the first word of the filetype, which he tests for equality.  I’m doing a pattern match, like you, so I could have just used the entire output, but I use cut to get the filetype only, without the filename.)  Also, Stéphane Chazelas's answer uses file < "$attachment", and, while not explained, I presume that that trick is also intended to get output without the filename. – G-Man Jun 12 '15 at 18:58
  • @G-Man Indeed that was a pretty big elephant. Thanks for your comment. I've updated making the regex way more consistent. – kos Jun 12 '15 at 20:42
  • 1
    @kos, now you're moving the problem to files with : in their name. – Stéphane Chazelas Jun 12 '15 at 20:46
  • 1
    @G-Man, I used file - < "$attachment". file - tries to identify the file content of stdin. so it's simliar to file -sL -- "$attachment" except that it also work for a file called - (-- doesn't help with that). – Stéphane Chazelas Jun 12 '15 at 22:23

Not the answer you're looking for? Browse other questions tagged or ask your own question.