shell script for reading file and replacing new file with | symbol

Question

i have txt file like below.

abc
def
ghi

123
456
789

expected output is

abc|def|ghi
123|456|789

I want replace new line with pipe symbol (|). i want to use in egrep.After empty line it should start other new line.

Answer 1

you can try with awk

awk -v RS= -v OFS="|" '{$1=$1}1' file

you get,

abc|def|ghi
123|456|789

Explanation

Set RS to a null/blank value to get awk to operate on sequences of blank lines.

From the POSIX specification for awk:

RS

The first character of the string value of RS shall be the input record separator; a by default. If RS contains more than one character, the results are unspecified. If RS is null, then records are separated by sequences consisting of a plus one or more blank lines, leading or trailing blank lines shall not result in empty records at the beginning or end of the input, and a shall always be a field separator, no matter what the value of FS is.

$1==$1 re-formatting output with OFS as separator, 1 is true for always print.

Answer 2

Here's one using GNU sed:

cat file | sed ':a; N; $!ba; s/\n/|/g; s/||/\n/g'

If you're using BSD sed (the flavor packaged with Mac OS X), you will need to pass in each expression separately, and use a literal newline instead of \n (more info):

cat file | sed -e ':a' -e 'N' -e '$!ba' -e 's/\n/|/g' -e 's/||/\
/g'

If file is:

abc
def
ghi

123
456
789

You get:

abc|def|ghi
123|456|789

This replaces each newline with a | (credit to this answer), and then || (i.e. what was a pair of newlines in the original input) with a newline.

The caveat here is that | can't appear at the beginning or end of a line in your input; otherwise, the second sed will add newlines in the wrong places. To work around that, you can use another character that won't be in your input as an intermediate value, and then replace singletons of that character with | and pairs with \n.

EDIT

Here's an example that implements the workaround above, using the NUL character \x00 (which should be highly unlikely to appear in your input) as the intermediate character:

cat file | sed ':a;N;$!ba; s/\n/\x00/g; s/\x00\x00/\n/g; s/\x00/|/g'

Explanation:

• :a;N;$!ba; puts the entire file in the pattern space, including newlines
• s/\n/\x00/g; replaces all newlines with the NUL character
• s/\x00\x00/\n/g; replaces all pairs of NULs with a newline
• s/\x00/|/g replaces the remaining singletons of NULs with a |

BSD version:

sed -e ':a' -e 'N' -e '$!ba' -e 's/\n/\x00/g' -e 's/\x00\x00/\
/g' -e 's/\x00/|/g'

EDIT 2

For a more direct approach (GNU sed only), provided by @ClaudiuGeorgiu:

sed -z 's/\([^\n]\)\n\([^\n]\)/\1|\2/g; s/\n\n/\n/g'

Explanation:

• -z uses NUL characters as line-endings (so newlines are not given special treatment and can be matched in the regular expression)
• s/\([^\n]\)\n\([^\n]\)/\1|\2/g; replaces every 3-character sequence of <non-newline><newline><non-newline> with <non-newline>|<non-newline>
• s/\n\n/\n/g replaces all pairs of newlines with a single newline

Answer 3

In native bash:

#!/usr/bin/env bash
curr=
while IFS= read -r line; do
  if [[ $line ]]; then
    curr+="|$line"
  else
    printf '%s\n' "${curr#|}"
    curr=
  fi
done
[[ $curr ]] && printf '%s\n' "${curr#|}"

Tested:

$ f() { local curr= line; while IFS= read -r line; do if [[ $line ]]; then curr+="|$line"; else printf '%s\n' "${curr#|}"; curr=; fi; done; [[ $curr ]] && printf '%s\n' "${curr#|}"; }
$ f < <(printf '%s\n' 'abc' 'def' 'ghi' '' 123 456 789)
abc|def|ghi
123|456|789

Answer 4

Use rs. For example:

rs -C'|' 2 3 < file

rs = reshape data array. Here I'm specifying that I want 2 rows, 3 columns, and the output separator to be pipe.