Bash checking to see if a line only has a certain characters

Question

i am wanting to check that the first line of a file to see if it contains only "X", if it contains somthing other then "X" on the first line then i want to echo "yes" if not echo "no" for example:

the first line in a file is XXXXXXX i need the output to be yes

additionally how can i do this with the last line as well ?

Answer 1

The following script takes a filename as a parameter and checks if the first and the last line are composed of only X characters

#!/bin/bash
echo First line
if [ "$(head -n 1 $1 | grep '^X+$')" == "" ]; then
    #echo something else than X found in the first line
    echo no
else
    #echo only X found in the first line
    echo yes
fi
echo Last line
if [ "$(tail -n 1 $1 | grep '^X+$')" == "" ]; then
    #echo something else than X found in the last line
    echo no
else
    #echo only X found in the last line
    echo yes
fi

Answer 2

Because of the title ("… if a line only has a certain characters") I assume "only X" means "only X characters", not strictly "only one X character".

With sed. Let me start with checking the last line because it's more general:

sed -n -e '$ ! d' -e '/^X*$/ ino' -e '/^X*$/ ! iyes' -e '$ q'

To check the N-th line replace $ in the first and in the last expression with N. E.g. for N=5:

sed -n -e '5 ! d' -e '/^X*$/ ino' -e '/^X*$/ ! iyes' -e '5 q'

N=1 can be simplified because in this case we don't need the first expression and the last one can be just q:

sed -n -e '/^X*$/ ino' -e '/^X*$/ ! iyes' -e q

Notes:

• If there is no N-th line then there will be no output.

• Thanks to q sed exits as soon as possible (lines after N-th are not processed).

• You can use ^XX*$ instead of ^X*$. These will generate different output if the line in question happens to be empty. To test for a line being exactly X use ^X$.

• In the general case N appears twice and ^X*$ appears twice. It's not DRY but you can use shell variables:

  N='$'
  pattern='^X*$'
  sed -n -e "$N ! d" -e "/$pattern/ ino" -e "/$pattern/ ! iyes" -e "$N q"
  

Answer 3

In a similar way as in my previous answers to your questions (here and here): Delete all other lines and replace the current line with the string yes or no depending on whether it contains any other character than a single X.

sed '1!d; s/^X$/no/; t; s/.*/yes/' file

The pattern ^X$ matches a lone X on a line, so the first substitution would replace the whole line with the string no if there is only a single X character on it. The t command by itself skips the rest of the script if the most recent s/// command made a change.

To look at the last line, use $ in place of 1 in the sed code:

sed '$!d; s/^X$/no/; t; s/.*/yes/' file

Answer 4

Using AWK checking the first line:

awk '{ print $0!="X"?"yes":"no"; exit }' infile

same for checking the last line:

awk 'END{ print $0!="X"?"yes":"no" }' infile

and checking on N^th line (replace N in NR==N with the line-number you want verify it is only X or not; NR==line#):

awk 'NR==N{ print $0!="X"?"yes":"no"; exit }' infile

Answer 5

If you want to also consider lines that may contain NUL characters (or more generally non-text, which also includes sequence of bytes not forming valid characters or lines longer than LINE_MAX or lines not delimited by a newline character), you may need to switch to zsh as other shells can't cope with that easily, and text utilities are not guaranteed to either:

#! /bin/zsh -
if IFS= read -r line < file; then
  case $line in
    ("")     print -u2 empty;
    (*[^X]*) print -u2 does not contain only Xes;;
    (*)      print OK, only Xes;;
  esac
else
  print -u2 a full lines could not be read
fi

Or resort to perl:

perl -e '
  open FILE, "<", "file" or die "Cannot open: $!\n";;
  $_ = <FILE>;
  die "a line could not be read\n" unless defined $_;
  die "a full line could not be read\n" unless /\n$/;
  chomp;
  die "does not contain only Xes\n" if /[^X]/;
  print "OK, only Xes\n"'