3

I have a text file with following contents

$ cat foo.txt

some text 
[email protected]
8903457923
2018-02-09 07:12 (Asia/Kolkata)
again some text over here
some more text again
Message
some text 
[email protected]
8903457923
2018-02-05 07:12 (Asia/Kolkata)
again some text over here
some more text again
Message

I would like to get following output

$ cat foo.txt

some text  [email protected]  8903457923  2018-02-09 07:12 (Asia/Kolkata)  again some text over her  some more text again  Message

some text [email protected]  8903457923  2018-02-05 07:12 (Asia/Kolkata) again some text over here  some more text again  Message

I guess I can achive this using tr and taking "Message" as a common string. But not sure how to implement this.

Improve this question
2
  • is the blank line between the two output data lines important? Commented Jun 20, 2018 at 15:06
  • @JeffSchaller yes Commented Jun 20, 2018 at 15:10

5 Answers 5

Reset to default
6

If the current line is not "Message", then append the line to the list, joined with OFS; when you see "Message", print the current list (joined by OFS with the current "Message" line):

awk '/^Message$/ { print t OFS $0 ORS; t=""; } !/^Message$/ { t=(t ? t OFS $0 : $0) }' < foo.txt

The t=(t ? t OFS $0 : $0) part is a ternary operator; it checks to see if t is empty; if it is, then just assign the current line to it; otherwise, append the current value with OFS followed by the current line.

Output:

some text  [email protected] 8903457923 2018-02-09 07:12 (Asia/Kolkata) again some text over here some more text again Message

some text  [email protected] 8903457923 2018-02-05 07:12 (Asia/Kolkata) again some text over here some more text again Message
Improve this answer
0
5

a bit easier way with AWK:

awk 'BEGIN { ORS=RS="Message\n" } gsub("\n"," ")' ./in.txt

Improve this answer
1
  • note that not all versions support multi-character RS.. Commented Jun 21, 2018 at 4:10
2

Using tr and sed:

tr '\n' '\t' <foo.txt  | sed -e $'s/Message\\\t/Message\\\n\\\n/g'

This will convert all newlines to tabs and then add a two newlines after each occurrence of Message

Improve this answer
0

You can try this sed

sed ':A;/Message/!{N;bA};y/\n/ /;$!G' foo.txt

Get lines in pattern space until you see Message.
Replace each newline by blank space.
Add a blank line only if it's not the last.

Improve this answer
0

The simplest and most direct method is :

 perl -lpe '$\ = /^Message$/ ? "\n\n" : " "' foo.txt

or using awk it would be:

 awk 'ORS = /^Message$/ ? RS RS : " "'

Slurp the file, -0777 then the whole file is a one large string upon which the regex /(.*?)^(Message\n)/msg is operated. The regex looks at the shortest chunk which abuts the Message line and stores the chunk in $1 and the nessage line in $2. The chunk comprises newlines which are replaced globally by space and the result of this transformation tr/\n/ /r is delivered to print. The while loop keeps going so long as the chunk + message lines are to be found.

perl -ln -0777e 'print $1 =~ tr/\n/ /r, $2 while /(.*?)^(Message\n)/msg' foo.txt

We could use the sed tool to perform this job as shown here:

sed -e '
    $!N;G;s/\n/ /                ;# put 2 lines in pattern space
    / Message\n$/b               ;# one message block has been found
    s/\(.*\)\(.\)/\2\1/;D        ;# go back to read the next line into pattern space
' foo.txt
Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.