I want to find if a word exists in a file

Question

I have a directory with some files. I want to know if a word exists in each file. $1 is my directory that I give it as a parameter

e.g. ./myscript.sh /mnt/c/User/er/Desktop/shows

I have written this code but the terminal shows this error : [: it21754 :binary operator expected

#!/bin/bash
for a in $( find "$1" -type f )
do
      if[ egrep 'it21754' $a ]
      then 
          echo "0" 
      else 
          echo "1"
      fi 
done

Answer 1

Don't loop over the output of find. It's rather inelegant as the complete find command has to finish and produce all its output before the loop can even run its first iteration. Furthermore you disqualify your script from working correctly with filenames that contain spaces. The result of the command substitution would be split on spaces to produce words for the loop to loop over.

Instead, let the find command generate filenames for the loop like this:

find "$1" -type f -exec sh -c '...loop here...' sh {} +

We'll come back to this shortly.

The syntax on the line

if[ egrep 'it21754' $a ]

is a bit wonky, and you definitely do not need egrep here (grep is enough).

By "wonky" I mean that a [ ... ] test usually looks like [ op arg ] or [ arg op arg ] where op is some operator like -f or -eq or = and arg is a string. You use egrep and it21754 and whatever $a expands to, which is what generates the error that you get. You can't run egrep in this context really, and you don't need to.

The correct if-statement is

if grep -qwF 'it21754' "$a"; then
    echo 0
else
    echo 1
fi

The -qwF flags makes grep quiet (-q), it makes it match using string comparisons rather than regular expressions (-F), and it assures that the string matched will be a complete word rather than a substring (-w). Here we use the exit status of grep to choose whether to output a 0 (for a match) or 1 (for no match).

I've also double quoted $a as "$a" just in case the filename contains any spaces or filename globbing characters.

Interestingly enough, those are the numbers that grep uses for its exit status anyway, so we could shorten the whole thing to just

grep -qwF 'it21754' "$a"
echo "$?"

The $? variable holds the exit status of the most recently executed command.

Plugging that into our find command, we get

find "$1" -type f -exec sh -c '
    for a do
        grep -qwF "it21754" "$a"
        echo "$?"
    done' sh {} +

This will produce a number of ones and zeroes depending on whether the files contain the given word or not. However, I think you may be looking for the filenames of the files that contain the word rather than a zero or a one. For that we may use

find "$1" -type f -exec grep -lwF 'it21754' {} +

Here, I've changed the -q to a -l ("dash ell"). This calls grep with the found files and grep will only output the filenames of the ones that contains the given word. We don't need the sh -c bit any longer since we're calling a single utility (not grep and echo).

Of course, you could do the same thing with just grep:

grep -R -lwF 'it21754' "$1"

With -R, grep will search the subdirectories given on the command line recursively.

It may be worth noting that the -w and -R flags of grep are not standard, but implemented by the most common grep variants.

Related:

• Understanding the -exec option of `find`