13

Is there a method to print multi-line output (single output) on the same line?

For example, if the output is:

abc
def
qwerty

Is it possible to print:

abcdefqwerty 
12

You can remove all occurrences of characters from a given set with tr -d. To remove the newline character use:

tr -d '\n'

As always you can use input and output redirection and pipes to read from or write to files and other processes.

If you want to keep the last newline you can simply add it back with echo or printf '\n', e. g.:

cat file1 file2... | { tr -d '\n'; echo; } > output.txt
10

Many ways. To illustrate, I have saved your example in file:

$ cat file | tr -d '\n'
abcdefqwerty$ 
$ cat file | perl -pe 's/\n//'
abcdefqwerty$ 

That removes all newline characters, including the last one though. You might instead want to do:

$ printf "%s\n" "$(cat file | perl -pe 's/\n//')"
abcdefqwerty
$ printf "%s\n" "$(cat file | tr -d '\n')"
abcdefqwerty
  • I think the printf + cat + perl combo is much better handled via perl -pe 's/\n//;END{printf "\n"}' input.txt Single process, no command-substitution and messing with quotes, and no UUOC. Perhaps. – Sergiy Kolodyazhnyy Jun 1 '18 at 0:10
7

You can pipe your multiline output through awk

awk '{printf "%s",$0} END {print ""}'

or use sed:

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

Example run

$ echo -e "1\n2\n3\n4" | awk '{printf "%s",$0} END {print ""}'
1234
$ echo -e "1\n2\n3\n4" | sed ':a;N;$!ba;s/\n//g'              
1234

Further reading: remove line break using AWK · serverfault.SE

4

Also your_command | paste -sd '' which preserves the trailing newline.

2

If you want to use Perl for this, you can use its chomp function:

perl -pe chomp

Some more details, for those who are interested.

You can pass one or more filenames as subsequent arguments.

perl -pe chomp file1 file2 file3

In some cases you might not want to do that, but you can pipe or redirect to that command instead. (These abilities--and that caveat--all apply to any other perl -p or perl -n based solution, too.)

So, if you want to process output from running some-command:

some-command | perl -pe chomp

This is potentially faster than the Perl command in terdon's answer. Newline characters (represented by \n) only appear at the end of lines in this situation--because they are what Perl is using to decide where each line ends. s/\n// searches through the entire line for newlines, while chomp just removes the one at the end (if any, as the very last line might or might not have one).

Performance might not really be the main reason to prefer chomp. The perl command in terdon's answer would only be slightly slower unless you're processing a huge file with very long lines. And if you need speed, David Foerster's way is still probably faster. chomp is, however, exactly the right tool for this if you are using Perl, and I think the intent of chomp is clearer than that of s/\n//. Ultimately there's probably no objective answer as to which is better.

Printing a final newline.

However, I do recommend against the use of command substitution ($( )) just to ensure a trailing newline gets printed. Unless the text you're processing is short, that's likely to be quite slow--it has to collect all the text at once and then print it--and it makes your command harder to understand. Instead you can do as David Foerster suggested and run echo or printf '\n' afterwards.

perl -pe chomp; echo

If you're piping from or (as David Foerster shows) redirecting from that command, then you can enclose it in { ;} so that the output of both commands is taken together. If you're just printing it to the terminal, then you can run it exactly as shown above. This works even if you're piping or redirecting to the command, because echo/printf '\n' does not actually read any input. That is, this works:

some-command | perl -pe chomp; echo

While here you cannot simply remove the { ;}:

{ perl -pe chomp; echo; } | some-other-command

Or, if you prefer, you can make perl print the final newline itself. Like enclosing the perl and echo commands in { ;}, this approach works even if you're piping or redirecting from it (and, like all approaches, this works when you're piping to it, too):

perl -wpe 'chomp; END { print "\n" }'

Note that the command in terdon's answer works fine with these methods of printing the final newline, too. For example:

perl -pe 's/\n//'; echo
perl -pe 's/\n//; END { print "\n" }'
  • perl -wpe 'chomp; END { print "\n" } is same process, minor advantage over adding trailing echo – Sergiy Kolodyazhnyy Mar 17 '18 at 18:18
1

Using shell-only way:

while IFS= read -r line || [ -n "$line" ]; do printf "%s"  "$line"; done < inputfile.txt
1

This answer has a solution to the problem you are trying to create: Why does bash remove \n in $(cat file)?

If you type cat myfile.txt you will see:

abc
def
ghi

But if you type echo $(cat myfile.txt) you will see:

abc def ghi

Note this method inserts a space where separate new lines used to be. This makes the output easier to read but doesn't strictly adhere to your question scope.

0

If you are using Bash, then you can do it without any external tool.

x=($(echo line1; echo line2))
echo "${x[@]}"

Result:

line1 line2

The first command turn multiline output into an array, the second command expand array into multiple arguments on one line.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.