Optional option argument with getopts

Question

while getopts "hd:R:" arg; do
  case $arg in
    h)
      echo "usgae" 
      ;;
    d)
      dir=$OPTARG
      ;;
    R)
      if [[ $OPTARG =~ ^[0-9]+$ ]];then
        level=$OPTARG
      else
        level=1
      fi
      ;;
    \?)
      echo "WRONG" >&2
      ;;
  esac
done

• level refers to parameter of -R, dir refers to parameters of -d

• when I input ./count.sh -R 1 -d test/ it works rightly

• when I input ./count.sh -d test/ -R 1 it works rightly

• but I want to make it work when I input ./count.sh -d test/ -R or ./count.sh -R -d test/

This means that I want -R has a default value and the sequence of command could be more flexible.

Answer 1

The following code solves this problem by checking for a leading dash and if found decrements OPTIND to point back to the skipped option for processing. This generally works fine except that you do not know the order the user will place options on the command line - if your optional argument option is last and does not provide an argument getopts will want to error out.

To fix the problem of the final argument missing, the "$@" array simply has an empty string "$@ " appended so that getopts will be satisfied that it has gobbled up yet another option argument. To fix this new empty argument a variable is set that holds the total count of all options to be processed - when the last option is being processed a helper function called trim is called and removes the empty string prior to the value being utilized.

This is not working code, it has only place holders but you can easily modify it and with a little bit of care it can be useful to build a robust system.

#!/usr/bin/env bash 
declare  -r CHECK_FLOAT="%f"  
declare  -r CHECK_INTEGER="%i"  

 ## <arg 1> Number - Number to check
 ## <arg 2> String - Number type to check
 ## <arg 3> String - Error message
function check_number() {
  local NUMBER="${1}"
  local NUMBER_TYPE="${2}"
  local ERROR_MESG="${3}"
  local FILTERED_NUMBER=$(sed 's/[^.e0-9+\^]//g' <<< "${NUMBER}")
  local -i PASS=1
  local -i FAIL=0
    if [[ -z "${NUMBER}" ]]; then 
        echo "Empty number argument passed to check_number()." 1>&2
        echo "${ERROR_MESG}" 1>&2
        echo "${FAIL}"          
  elif [[ -z "${NUMBER_TYPE}" ]]; then 
        echo "Empty number type argument passed to check_number()." 1>&2
        echo "${ERROR_MESG}" 1>&2
        echo "${FAIL}"          
  elif [[ ! "${#NUMBER}" -eq "${#FILTERED_NUMBER}" ]]; then 
        echo "Non numeric characters found in number argument passed to check_number()." 1>&2
        echo "${ERROR_MESG}" 1>&2
        echo "${FAIL}"          
  else  
   case "${NUMBER_TYPE}" in
     "${CHECK_FLOAT}")
         if ((! $(printf "${CHECK_FLOAT}" "${NUMBER}" &>/dev/random;echo $?))); then
            echo "${PASS}"
         else
            echo "${ERROR_MESG}" 1>&2
            echo "${FAIL}"
         fi
         ;;
     "${CHECK_INTEGER}")
         if ((! $(printf "${CHECK_INTEGER}" "${NUMBER}" &>/dev/random;echo $?))); then
            echo "${PASS}"
         else
            echo "${ERROR_MESG}" 1>&2
            echo "${FAIL}"
         fi
         ;;
                      *)
         echo "Invalid number type format: ${NUMBER_TYPE} to check_number()." 1>&2
         echo "${FAIL}"
         ;;
    esac
 fi 
}

 ## Note: Number can be any printf acceptable format and includes leading quotes and quotations, 
 ##       and anything else that corresponds to the POSIX specification. 
 ##       E.g. "'1e+03" is valid POSIX float format, see http://mywiki.wooledge.org/BashFAQ/054
 ## <arg 1> Number - Number to print
 ## <arg 2> String - Number type to print
function print_number() { 
  local NUMBER="${1}" 
  local NUMBER_TYPE="${2}" 
  case "${NUMBER_TYPE}" in 
      "${CHECK_FLOAT}") 
           printf "${CHECK_FLOAT}" "${NUMBER}" || echo "Error printing Float in print_number()." 1>&2
        ;;                 
    "${CHECK_INTEGER}") 
           printf "${CHECK_INTEGER}" "${NUMBER}" || echo "Error printing Integer in print_number()." 1>&2
        ;;                 
                     *) 
        echo "Invalid number type format: ${NUMBER_TYPE} to print_number()." 1>&2
        ;;                 
   esac
} 

 ## <arg 1> String - String to trim single ending whitespace from
function trim_string() { 
 local STRING="${1}" 
 echo -En $(sed 's/ $//' <<< "${STRING}") || echo "Error in trim_string() expected a sensible string, found: ${STRING}" 1>&2
} 

 ## This a hack for getopts because getopts does not support optional
 ## arguments very intuitively. E.g. Regardless of whether the values
 ## begin with a dash, getopts presumes that anything following an
 ## option that takes an option argument is the option argument. To fix  
 ## this the index variable OPTIND is decremented so it points back to  
 ## the otherwise skipped value in the array option argument. This works
 ## except for when the missing argument is on the end of the list,
 ## in this case getopts will not have anything to gobble as an
 ## argument to the option and will want to error out. To avoid this an
 ## empty string is appended to the argument array, yet in so doing
 ## care must be taken to manage this added empty string appropriately.
 ## As a result any option that doesn't exit at the time its processed
 ## needs to be made to accept an argument, otherwise you will never
 ## know if the option will be the last option sent thus having an empty
 ## string attached and causing it to land in the default handler.
function process_options() {
local OPTIND OPTERR=0 OPTARG OPTION h d r s M R S D
local ERROR_MSG=""  
local OPTION_VAL=""
local EXIT_VALUE=0
local -i NUM_OPTIONS
let NUM_OPTIONS=${#@}+1
while getopts “:h?d:DM:R:S:s:r:” OPTION "$@";
 do
     case "$OPTION" in
         h)
             help | more
             exit 0
             ;;
         r)
             OPTION_VAL=$(((${NUM_OPTIONS}==${OPTIND})) && trim_string "${OPTARG##*=}" || echo -En "${OPTARG##*=}")
             ERROR_MSG="Invalid input: Integer or floating point number required."
             if [[ -z "${OPTION_VAL}" ]]; then
               ## can set global flags here 
               :;
             elif [[ "${OPTION_VAL}" =~ ^-. ]]; then
               let OPTIND=${OPTIND}-1
               ## can set global flags here 
             elif [ "${OPTION_VAL}" = "0" ]; then
               ## can set global flags here 
               :;               
             elif (($(check_number "${OPTION_VAL}" "${CHECK_FLOAT}" "${ERROR_MSG}"))); then
               :; ## do something really useful here..               
             else
               echo "${ERROR_MSG}" 1>&2 && exit -1
             fi
             ;;
         d)
             OPTION_VAL=$(((${NUM_OPTIONS}==${OPTIND})) && trim_string "${OPTARG##*=}" || echo -En "${OPTARG##*=}")
             [[  ! -z "${OPTION_VAL}" && "${OPTION_VAL}" =~ ^-. ]] && let OPTIND=${OPTIND}-1            
             DEBUGMODE=1
             set -xuo pipefail
             ;;
         s)
             OPTION_VAL=$(((${NUM_OPTIONS}==${OPTIND})) && trim_string "${OPTARG##*=}" || echo -En "${OPTARG##*=}")
             if [[ ! -z "${OPTION_VAL}" && "${OPTION_VAL}" =~ ^-. ]]; then ## if you want a variable value that begins with a dash, escape it
               let OPTIND=${OPTIND}-1
             else
              GLOBAL_SCRIPT_VAR="${OPTION_VAL}"
                :; ## do more important things
             fi
             ;;
         M)  
             OPTION_VAL=$(((${NUM_OPTIONS}==${OPTIND})) && trim_string "${OPTARG##*=}" || echo -En "${OPTARG##*=}")
             ERROR_MSG=$(echo "Error - Invalid input: ${OPTION_VAL}, Integer required"\
                              "retry with an appropriate option argument.")
             if [[ -z "${OPTION_VAL}" ]]; then
               echo "${ERROR_MSG}" 1>&2 && exit -1
             elif [[ "${OPTION_VAL}" =~ ^-. ]]; then
               let OPTIND=${OPTIND}-1
               echo "${ERROR_MSG}" 1>&2 && exit -1
             elif (($(check_number "${OPTION_VAL}" "${CHECK_INTEGER}" "${ERROR_MSG}"))); then
             :; ## do something useful here
             else
               echo "${ERROR_MSG}" 1>&2 && exit -1
             fi
             ;;                      
         R)  
             OPTION_VAL=$(((${NUM_OPTIONS}==${OPTIND})) && trim_string "${OPTARG##*=}" || echo -En "${OPTARG##*=}")
             ERROR_MSG=$(echo "Error - Invalid option argument: ${OPTION_VAL},"\
                              "the value supplied to -R is expected to be a "\
                              "qualified path to a random character device.")            
             if [[ -z "${OPTION_VAL}" ]]; then
               echo "${ERROR_MSG}" 1>&2 && exit -1
             elif [[ "${OPTION_VAL}" =~ ^-. ]]; then
               let OPTIND=${OPTIND}-1
               echo "${ERROR_MSG}" 1>&2 && exit -1
             elif [[ -c "${OPTION_VAL}" ]]; then
               :; ## Instead of erroring do something useful here..  
             else
               echo "${ERROR_MSG}" 1>&2 && exit -1
             fi
             ;;                      
         S)  
             STATEMENT=$(((${NUM_OPTIONS}==${OPTIND})) && trim_string "${OPTARG##*=}" || echo -En "${OPTARG##*=}")
             ERROR_MSG="Error - Default text string to set cannot be empty."
             if [[ -z "${STATEMENT}" ]]; then
               ## Instead of erroring you could set a flag or do something else with your code here..  
             elif [[ "${STATEMENT}" =~ ^-. ]]; then ## if you want a statement that begins with a dash, escape it
               let OPTIND=${OPTIND}-1
               echo "${ERROR_MSG}" 1>&2 && exit -1
               echo "${ERROR_MSG}" 1>&2 && exit -1
             else
                :; ## do something even more useful here you can modify the above as well 
             fi
             ;;                      
         D)  
             ## Do something useful as long as it is an exit, it is okay to not worry about the option arguments 
             exit 0
             ;;          
         *)
             EXIT_VALUE=-1
             ;&                  
         ?)
             usage
             exit ${EXIT_VALUE}
             ;;
     esac
done
}

process_options "$@ " ## extra space, so getopts can find arguments  

Answer 2

Try:

while getopts "hd:R:" arg; do
  case $arg in
    h)
      echo "usage" 
    ;;
    d)
      dir=$OPTARG
    ;;
    R)
      if [[ $OPTARG =~ ^[0-9]+$ ]];then
        level=$OPTARG
      elif [[ $OPTARG =~ ^-. ]];then
        level=1
        let OPTIND=$OPTIND-1
      else
        level=1
      fi          
    ;;
    \?)
      echo "WRONG" >&2
    ;;
  esac
done

I think the above code will work for your purposes while still using getopts. I've added the following three lines to your code when getopts encounters -R:

      elif [[ $OPTARG =~ ^-. ]];then
        level=1
        let OPTIND=$OPTIND-1

If -R is encountered and the first argument looks like another getopts parameter, level is set to the default value of 1, and then the $OPTIND variable is reduced by one. The next time getopts goes to grab an argument, it will grab the correct argument instead of skipping it.

---

Here is similar example based on the code from Jan Schampera's comment at this tutorial:

#!/bin/bash
while getopts :abc: opt; do
  case $opt in
    a)
      echo "option a"
    ;;
    b)
      echo "option b"
    ;;
    c)
      echo "option c"

      if [[ $OPTARG = -* ]]; then
        ((OPTIND--))
        continue
      fi

      echo "(c) argument $OPTARG"
    ;;
    \?)
      echo "WTF!"
      exit 1
    ;;
  esac
done

When you discover that OPTARG von -c is something beginning with a hyphen, then reset OPTIND and re-run getopts (continue the while loop). Oh, of course, this isn't perfect and needs some more robustness. It's just an example.

Answer 3

Wrong. Actually getopts does support optional arguments! From the bash man page:

If  a  required  argument is not found, and getopts is not silent, 
a question mark (?) is placed in name, OPTARG is unset, and a diagnostic
message is printed.  If getopts is silent, then a colon (:) is placed in name 
and OPTARG is set to the option character found.

When the man page says "silent" it means silent error reporting. To enable it, the first character of optstring needs to be a colon:

while getopts ":hd:R:" arg; do
    # ...rest of iverson's loop should work as posted 
done

Since Bash's getopt does not recognize -- to end the options list, it may not work when -R is the last option, followed by some path argument.

P.S.: Traditionally, getopt.c uses two colons (::) to specify an optional argument. However, the version used by Bash doesn't.

Answer 4

I agree with tripleee, getopts does not support optional argument handling.

The compromised solution I have settled on is to use the upper case/lower case combination of the same option flag to differentiate between the option that takes an argument and the other that does not.

Example:

COMMAND_LINE_OPTIONS_HELP='
Command line options:
    -I          Process all the files in the default dir: '`pwd`'/input/
    -i  DIR     Process all the files in the user specified input dir
    -h          Print this help menu

Examples:
    Process all files in the default input dir
        '`basename $0`' -I

    Process all files in the user specified input dir
        '`basename $0`' -i ~/my/input/dir

'

VALID_COMMAND_LINE_OPTIONS="i:Ih"
INPUT_DIR=

while getopts $VALID_COMMAND_LINE_OPTIONS options; do
    #echo "option is " $options
    case $options in
        h)
            echo "$COMMAND_LINE_OPTIONS_HELP"
            exit $E_OPTERROR;
        ;;
        I)
            INPUT_DIR=`pwd`/input
            echo ""
            echo "***************************"
            echo "Use DEFAULT input dir : $INPUT_DIR"
            echo "***************************"
        ;;
        i)
            INPUT_DIR=$OPTARG
            echo ""
            echo "***************************"
            echo "Use USER SPECIFIED input dir : $INPUT_DIR"
            echo "***************************"
        ;;
        \?)
            echo "Usage: `basename $0` -h for help";
            echo "$COMMAND_LINE_OPTIONS_HELP"
            exit $E_OPTERROR;
        ;;
    esac
done

Answer 5

This is actually pretty easy. Just drop the trailing colon after the R and use OPTIND

while getopts "hRd:" opt; do
   case $opt in
      h) echo -e $USAGE && exit
      ;;
      d) DIR="$OPTARG"
      ;;
      R)       
        if [[ ${@:$OPTIND} =~ ^[0-9]+$ ]];then
          LEVEL=${@:$OPTIND}
          OPTIND=$((OPTIND+1))
        else
          LEVEL=1
        fi
      ;;
      \?) echo "Invalid option -$OPTARG" >&2
      ;;
   esac
done
echo $LEVEL $DIR

count.sh -d test

test

count.sh -d test -R

1 test

count.sh -R -d test

1 test

count.sh -d test -R 2

2 test

count.sh -R 2 -d test

2 test

Answer 6

You can always decide to differentiate the option with lowercase or uppercase.

However my idea is to call getopts twice and 1st time parse without arguments ignoring them (R) then 2nd time parse only that option with argument support (R:). The only trick is that OPTIND (index) needs to be changed during processing, as it keeps pointer to the current argument.

Here is the code:

#!/usr/bin/env bash
while getopts ":hd:R" arg; do
  case $arg in
    d) # Set directory, e.g. -d /foo
      dir=$OPTARG
      ;;
    R) # Optional level value, e.g. -R 123
      OI=$OPTIND # Backup old value.
      ((OPTIND--)) # Decrease argument index, to parse -R again.
      while getopts ":R:" r; do
        case $r in
          R)
            # Check if value is in numeric format.
            if [[ $OPTARG =~ ^[0-9]+$ ]]; then
              level=$OPTARG
            else
              level=1
            fi
          ;;
          :)
            # Missing -R value.
            level=1
          ;;
        esac
      done
      [ -z "$level" ] && level=1 # If value not found, set to 1.
      OPTIND=$OI # Restore old value.
      ;;
    \? | h | *) # Display help.
      echo "$0 usage:" && grep " .)\ #" $0
      exit 0
      ;;
  esac
done
echo Dir: $dir
echo Level: $level

Here are few tests for scenarios which works:

$ ./getopts.sh -h
./getopts.sh usage:
    d) # Set directory, e.g. -d /foo
    R) # Optional level value, e.g. -R 123
    \? | h | *) # Display help.
$ ./getopts.sh -d /foo
Dir: /foo
Level:
$ ./getopts.sh -d /foo -R
Dir: /foo
Level: 1
$ ./getopts.sh -d /foo -R 123
Dir: /foo
Level: 123
$ ./getopts.sh -d /foo -R wtf
Dir: /foo
Level: 1
$ ./getopts.sh -R -d /foo
Dir: /foo
Level: 1

Scenarios which doesn't work (so the code needs a bit of more tweaks):

$ ./getopts.sh -R 123 -d /foo
Dir:
Level: 123

More information about getopts usage can be found in man bash.

See also: Small getopts tutorial at Bash Hackers Wiki