sed command to arrange words

Shell Programming and Scripting
1

Hi

I have a text file as given below:

Input format

I
have
a
text
file
as
given
below
.

This
is
a
sample
test
file
.

There is a space between the sentences in the above format.

I want to write sed command to arrange these words in such a way that they appear as given below.

Output Format:

I have a text file as given below .
This is a sample test file .

Thanks in advance. :slight_smile:

2 
sed -e ':L' -e '/\./{p;d;}' -e 'N;s/\n/ /;bL' file

label L
If there is . then p rint, d elete and next cycle.
Append N ext line (with embedded \n), s ubstitute the \n by a space, jump to label L
With some debug code you see how it works:

sed -e ':L' -e '/\./{p;d;}' -e '=;p;N;s/\n/ /;bL' file

print current line number and current pattern space.

1 Like
3

I got the following output:

I
have
a
text
file
as
given
below
.

This
is
a
sample
test
file
.
I have a text file as given below .
 This is a sample test file .

But I want to get the following output only:

I have a text file as given below .
This is a sample test file .
4 
sed -e ':L' -e '/\./{p;d;}' -e 's/^ //' -e 'N;s/\n/ /;bL'

Will do the trick!

---------- Post updated at 04:46 PM ---------- Previous update was at 04:44 PM ----------

Sorry, new here--

sed -e ':L' -e '/\./{p;d;}' -e 's/^ //' -e 'N;s/\n/ /;bL'

let's see if this is kosher!

1 Like
5

I got the following output:

I
have
a
text
file
as
given
below
.

This
is
a
sample
test
file
.
I have a text file as given below .
This is a sample test file .

How do I get only:

I have a text file as given below .
This is a sample test file .

Thanks.:slight_smile:

6 
sed -e ':L' -e '/\./{p;d;}' -e 's/^ //' -e 'N;s/\n/ /;bL'

Ah git it!!

1 Like
7

Or substitute the leading space only once before the print

sed -e ':L' -e '/\./{s/^ //;p;d;}' -e 'N;s/\n/ /;bL' file

You don't get the input repeated in the output. (What did you do?)

8

Not sed, but this awk alternative also works with your given input.

awk '{$1=$1}1' RS= file

Note: it only works if there are two consecutive newlines (a completely empty line) between two vertical sentences, as is the case in your sample.

9

Hi,
Another way in bash builtin works with your given input (for fun):

while read -d \. foo ; do echo $foo.; done <file ; [ "$foo" ] && echo $foo

Regards.

10

An alternative in Perl.

perl -00 -nae 'print "@F\n"' my_Perl.file