How to get the char at a given position of a string in shell script?

Question

score 59 · Accepted Answer · edited Aug 13 '21 at 17:06

59

In bash with "Parameter Expansion" ${parameter:offset:length}

$ var=abcdef
$ echo ${var:0:1}
a
$ echo ${var:3:1}
d

The same parameter expansion can be used to assign a new variable:

$ x=${var:1:1}
$ echo $x
b

Edit: Without parameter expansion (not very elegant, but that's what came to me first)

$ charpos() { pos=$1;shift; echo "$@"|sed 's/^.\{'$pos'\}\(.\).*$/\1/';}
$ charpos 8 what ever here
r

edited Aug 13 '21 at 17:06

d-man

103
4

answered Mar 17 '11 at 15:18

forcefsck

7,884
34
31

1

http://www.gnu.org/software/bash/manual/bashref.html#Shell-Parameter-Expansion has more details – jsbillings Mar 17 '11 at 15:24
Not POSIX apparently: http://pubs.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html – Ciro Santilli OurBigBook.com Aug 10 '17 at 14:03
1

You can also set the offset 'from end' like so `echo ${var: -2:1}` – Vassilis Jun 01 '18 at 15:37
That syntax comes from ksh93 and is also supported by `zsh` and `mksh`. – Stéphane Chazelas Sep 06 '19 at 08:08

score 12 · Answer 2 · answered Mar 17 '11 at 18:08

12

Alternative to parameter expansion is expr substr

substr STRING POS LENGTH
    substring of STRING, POS counted from 1

For example:

$ expr substr hello 2 1
e

answered Mar 17 '11 at 18:08

dogbane

29,087
16
80
60

cool, should have checked expr more thoroughly. – forcefsck Mar 17 '11 at 18:41
3

While this appears to work with the expr from GNU coreutils, `substr` is not included in the expr from FreeBSD, NetBSD or OS X. This isn't a portable solution. – ghoti Feb 14 '17 at 17:04
1

@ghoti, note that `substr` is not originally a GNU extension. The original implementation of `expr` came from PWB Unix in the late 70s and had `substr` (but not `:`). – Stéphane Chazelas Sep 06 '19 at 08:06
@StéphaneChazelas, thanks for adding historical perspective. :) While I'm pretty sure PWB usage isn't relevant to the OP, it's always fun to track features and changes through the decades. GNU tends to be many people's default, but in general, I think I'd avoid using options that aren't clearly [POSIX](https://pubs.opengroup.org/onlinepubs/9699919799/utilities/expr.html), and are known to be missing from major unices. – ghoti Sep 06 '19 at 20:47

score 11 · Answer 3 · edited Sep 06 '19 at 11:36

11

cut -c

If the variable does not contain newlines you can do:

myvar='abc'
printf '%s\n' "$myvar" | cut -c2

outputs:

awk substr is another POSIX alternative that works even if the variable has newlines:

myvar="$(printf 'a\nb\n')" # note that the last newline is stripped by
                           # the command substitution
awk -- 'BEGIN {print substr (ARGV[1], 3, 1)}' "$myvar"

outputs:

printf '%s\n' is to avoid problems with escape characters: https://stackoverflow.com/a/40423558/895245 e.g.:

myvar='\n'
printf '%s\n' "$myvar" | cut -c1

outputs \ as expected.

See also: https://stackoverflow.com/questions/1405611/extracting-first-two-characters-of-a-string-shell-scripting

Tested in Ubuntu 19.04.

edited Sep 06 '19 at 11:36

Stéphane Chazelas

522,931
91
1,010
1,501

answered Aug 07 '18 at 06:40

Ciro Santilli OurBigBook.com

17,176
4
113
99

1

`printf '%s' "$myvar" | cut -c2` is not POSIX as the output of `printf` is not text unless `$myvar` ends in a newline character. It otherwise assumes the variable doesn't contain newline characters as `cut` cuts *each line* of its input. – Stéphane Chazelas Sep 06 '19 at 08:11
1

The `awk` one would be more efficient and reliable with `awk -- 'BEGIN {print substr (ARGV[1], 2, 1)}' "$myvar"` – Stéphane Chazelas Sep 06 '19 at 08:11
Note that with current versions of GNU `cut`, that doesn't work for multi-byte characters (same for mawk or busybox awk) – Stéphane Chazelas Sep 06 '19 at 08:12
@StéphaneChazelas thanks for the points. I didn't understand what you mean in the first one: do you mean that `printf 'abc '| cut -c2` is wrong because no `\n` (this I don't know about), or that the command will fail if myvar has newlines (this I agree)? – Ciro Santilli OurBigBook.com Sep 06 '19 at 08:18
2

The behaviour of `cut` is unspecified if the input is not text (though `cut` implementations are required to handle lines or arbitrary length). The output of `printf abc` is not *text* as it doesn't end in a newline character. In practice depending on the implementation, if you pipe that to `cut -c2`, you get either `b`, `b` or nothing at all. You'd need `printf 'abc\n' | cut -c2` to get a behaviour specified by POSIX (that's required to output `b`) – Stéphane Chazelas Sep 06 '19 at 11:00
1

@StéphaneChazelas ah OK, awesome, I wasn't aware that POSIX defined what a "text file" is! https://unix.stackexchange.com/questions/446237/what-conditions-must-be-met-for-a-file-to-be-a-text-file-as-defined-by-posix – Ciro Santilli OurBigBook.com Sep 06 '19 at 11:02

score 2 · Answer 4 · answered Sep 06 '19 at 08:15

2

With zsh or yash, you'd use:

$ text='€$*₭£'
$ printf '%s\n' "${text[3]}"
*

(in zsh, you can shorten it to printf '%s\n' $text[3]).

answered Sep 06 '19 at 08:15

Stéphane Chazelas

522,931
91
1,010
1,501

score 1 · Answer 5 · answered May 17 '18 at 17:20

1

You can use the cut command. To get the 3rd postion:

echo "SAMPLETEXT" | cut -c3

Check this link http://www.folkstalk.com/2012/02/cut-command-in-unix-linux-examples.html

(Advanced cases) However, modifying IFS is also a good thing, especially when your input might have spaces. In that case alone, use the one below

saveifs=$IFS
IFS=$(echo -en "\n\b")
echo "SAMPLETEXT" | cut -c3
IFS=$saveifs

answered May 17 '18 at 17:20

Midhun Jose

33
6

1

I can't see how `IFS` would come into play in the code that you have posted. – Kusalananda Aug 07 '18 at 07:01

score 1 · Answer 6 · answered Dec 05 '21 at 16:02

This is a portable POSIX shell variant, which only uses builtins.
First a oneliner, than a better readable function.
It uses "parameter expansion" explained here:
https://pubs.opengroup.org/onlinepubs/9699919799/utilities/V3_chap02.html#tag_18_06_02

#!/bin/sh
x(){ s=$1;p=$2;i=0;l=${#s};while i=$((i+1));do r=${s#?};f=${s%"$r"};s=$r;case $i in $p)CHAR=$f&&return;;$l)return 1;;esac;done;}

x ABCDEF 3   # output substring at pos 3
echo $CHAR   # output is 'C'

Here the oneliner explained.

#!/bin/sh

string_get_pos()
{
  local string="$1"             # e.g. ABCDEF
  local pos="$2"                # e.g. 3
  local rest first i=0
  local length="${#string}"     # e.g. 6

  while i=$(( i + 1 )); do
    rest="${string#?}"          # e.g.  BCDEF
    first=${string%"$rest"}     # e.g. A
    string="$rest"

    case "$i" in
      $pos) export CHAR="$first" && return 0 ;;
      $length) return 1 ;;
    esac
  done
}

string_get_pos ABCDEF 3
echo $CHAR                # output is 'C'

score 0 · Answer 7 · answered Aug 08 '20 at 09:17

shell cut - print specific range of characters or given part from a string

#method1) using bash

 str=2020-08-08T07:40:00.000Z
 echo ${str:11:8}

#method2) using cut

 str=2020-08-08T07:40:00.000Z
 cut -c12-19 <<< $str

#method3) when working with awk

 str=2020-08-08T07:40:00.000Z
 awk '{time=gensub(/.{11}(.{8}).*/,"\\1","g",$1); print time}' <<< $str

How to get the char at a given position of a string in shell script?

7 Answers7