Test whether string is a valid integer

Question

[[ $var =~ ^-?[0-9]+$ ]]

• The ^ indicates the beginning of the input pattern
• The - is a literal "-"
• The ? means "0 or 1 of the preceding (-)"
• The + means "1 or more of the preceding ([0-9])"
• The $ indicates the end of the input pattern

So the regex matches an optional - (for the case of negative numbers), followed by one or more decimal digits.

References:

• http://www.tldp.org/LDP/abs/html/bashver3.html#REGEXMATCHREF

Wow... there are so many good solutions here!! Of all the solutions above, I agree with @nortally that using the -eq one liner is the coolest.

I am running GNU bash, version 4.1.5 (Debian). I have also checked this on ksh (SunSO 5.10).

Here is my version of checking if $1 is an integer or not:

if [ "$1" -eq "$1" ] 2>/dev/null
then
    echo "$1 is an integer !!"
else
    echo "ERROR: first parameter must be an integer."
    echo $USAGE
    exit 1
fi

This approach also accounts for negative numbers, which some of the other solutions will have a faulty negative result, and it will allow a prefix of "+" (e.g. +30) which obviously is an integer.

Results:

$ int_check.sh 123
123 is an integer !!

$ int_check.sh 123+
ERROR: first parameter must be an integer.

$ int_check.sh -123
-123 is an integer !!

$ int_check.sh +30
+30 is an integer !!

$ int_check.sh -123c
ERROR: first parameter must be an integer.

$ int_check.sh 123c
ERROR: first parameter must be an integer.

$ int_check.sh c123
ERROR: first parameter must be an integer.

The solution provided by Ignacio Vazquez-Abrams was also very neat (if you like regex) after it was explained. However, it does not handle positive numbers with the + prefix, but it can easily be fixed as below:

[[ $var =~ ^[-+]?[0-9]+$ ]]

Latecomer to the party here. I'm extremely surprised none of the answers mention the simplest, fastest, most portable solution; the case statement.

case ${variable#[-+]} in
  *[!0-9]* | '') echo Not a number ;;
  * ) echo Valid number ;;
esac

The trimming of any sign before the comparison feels like a bit of a hack, but that makes the expression for the case statement so much simpler.

I like the solution using the -eq test, because it's basically a one-liner.

My own solution was to use parameter expansion to throw away all the numerals and see if there was anything left. (I'm still using 3.0, haven't used [[ or expr before, but glad to meet them.)

if [ "${INPUT_STRING//[0-9]}" = "" ]; then
  # yes, natural number
else
  # no, has non-numeral chars
fi

For portability to pre-Bash 3.1 (when the =~ test was introduced), use expr.

if expr "$string" : '-\?[0-9]\+$' >/dev/null
then
  echo "String is a valid integer."
else
  echo "String is not a valid integer."
fi

expr STRING : REGEX searches for REGEX anchored at the start of STRING, echoing the first group (or length of match, if none) and returning success/failure. This is old regex syntax, hence the excess \. -\? means "maybe -", [0-9]\+ means "one or more digits", and $ means "end of string".

Bash also supports extended globs, though I don't recall from which version onwards.

shopt -s extglob
case "$string" of
    @(-|)[0-9]*([0-9]))
        echo "String is a valid integer." ;;
    *)
        echo "String is not a valid integer." ;;
esac

# equivalently, [[ $string = @(-|)[0-9]*([0-9])) ]]

@(-|) means "- or nothing", [0-9] means "digit", and *([0-9]) means "zero or more digits".

Here's yet another take on it (only using the test builtin command and its return code):

function is_int() { test "$@" -eq "$@" 2> /dev/null; } 
 
input="-123"
 
if is_int "$input"
then
   echo "Input: ${input}"
   echo "Integer: ${input}"
else
   echo "Not an integer: ${input}"
fi