BASH - grep - works on command line, but not script

Question

I am trying to get the number of occurrences in a script and when running the command on the command line, it works fine, but not in a script. Both variables are initialized. FILE_PATH is the absolute path of the file and VARIABLE_NAME is something like firstName or lastName.

VARIABLE_NAME="_firstName"    
grep -cP "\(this\.|\s\){1}$VARIABLE_NAME" $FILE_PATH

When I take $VARIABLE_NAME out of the script and replace it directly with what I am looking for, it works as expected, so it is something with the replacement.

Any ideas what I can try?

Answer 1

The problem seems to reside in your regex. can you show us what you are trying to match?

if you are trying to match constructs like this :

this.contentOfVar
  contentOfVar

try it with the following :

grep -cP "(this\.|\s)$VARIABLE_NAME" $FILE_PATH

This works in GNU grep. You don't have to specify {1}, since the default will be 1. With the -P flag on, \( and \) will match literal parenthesis in the string, since this is the perl behavior.

Answer 2

It looks like the problem is with regex escaping. Here is a complete working example of how I would do something like this.

#!/bin/bash

VARIABLE_NAMES=(testOne testTwo)
RE="(this\.|\s){1}"

cat > ./testfile.txt <<EOF
this.testOne
this.testTwo
this.testTwo
this.testOne
this.testTwo
EOF

for VARIABLE_NAME in ${VARIABLE_NAMES[@]}; do
    grep -cP "${RE}${VARIABLE_NAME}" testfile.txt
done

exit 0

Answer 3

To add to others' answers, I'd suggest using single-quotes except where you explicitly need something from a variable. For example:

#!/usr/bin/env bash

dosomething | grep -cP 'some stuff'${SOMEVARIABLE}'more stuff';

If you need single-quotes in the regex, it looks a little uglier:

# ...
dosomething | grep -cP 'some stuff'"'"'this part has a single-quote around it'"'"${SOMEVARIABLE}'more stuff';

The benefit of using single quotes is you don't have to escape things.

-Brian

Answer 4

It has to do with escaping the quotes and the slashes:

VARIABLE_NAME="_firstName"
FILE_PATH="/tmp/*"
echo grep -cP \"\\\(this\\\.\|\\\s\\\)${VARIABLE_NAME}\" $FILE_PATH
grep -cP \"\\\(this\\\.\|\\\s\\\)${VARIABLE_NAME}\" $FILE_PATH

I added the echo as an example to show how it's actually interpeted.