joining consecutive lines

[Franky Leeuwerck]

Hello,

I have file which looks like :
xxxxxxxxxx
xxx pattern xxx
yyy
xxx pattern xxx
yyy
xxxxxxxx

I would like to join consecutive lines so the output looks like :

xxxxxxxxxx
xxx pattern xxxyyy
xxx pattern xxxyyy
xxxxxxxx

How can this be realised with a simple sed or awk statement ?

Thanks in advance.
Franky

[Franky Leeuwerck]

Meanwhile I found a solution myself here it is.

sed '/pattern/N;s/\n/ /' filename

This will join only the lines with the pattern.

Thanks anyway.

Franky

[Peter Kloetgen]

Hi Franky,

you forgot to assign points to yourself for the quick solution of your problem....

:-)

Allways stay on the bright side of life!

Peter

[Franky Leeuwerck]

Peter,

Unfortunately you can not assign points for your own solutions. If that would have succeeded, then that would be my first points to be honest.

Regards,
Franky

[H.Merijn Brand (procura]

Give your opinion about my T-Shirt design, and I'll give you your first forum points. Deal?

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x3b617bb04b5cd611abdb0090277a778c,00.html

[Franky Leeuwerck]

Sorry, the solution was found too soon.

Seems that the problem is a bit more complicated. It is sometimes necessary to join 3 or more lines.

Suppose the file looks like :
xxxxx
xx patternA xx
patternB
xxxxx
xx patternA xx
xxxxx
patternB
xxxxx

The result file should like like :
xxxxx
xx patternA xxpatternB
xxxxx
xx patternA xxxxxxxpatternB
xxxxx

Any suggestions,

Thanks in advance
Franky

[Franky Leeuwerck]

Well,

I found the solution for the real problem, not on my last question :

cat file.org | sed -e '/startpattern/,/endpattern$/ d' > file.new

This will delete every block (multiple consecutive lines ) in the file starting with a line containing the startpattern and ending with a line ending with the endpattern.

Thanks anyway.

[Steven Sim Kok Leong]

Hi,

Below works on the assumption that "xx patternA" can itself be a static pattern. Apart from that, no problems getting everything from patternA to patternB to be on a the same line:

bash-2.05a# cat tstfile
xxxxx
xx patternA xx
patternB
xxxxx
xx patternA xx
xxxxx
patternB
xxxxx

bash-2.05a# echo `cat tstfile` | sed "s/ xx patternA/;xx patternA/g" | sed "s/patternB /patternB;/g" | tr [";"] ["\n"]
xxxxx
xx patternA xx patternB
xxxxx
xx patternA xx xxxxx patternB
xxxxx

Hope this helps. Regards.

Steven Sim Kok Leong

[H.Merijn Brand (procura]

Is it possible to reverse the problem?

Suppose the file looks like :
xxxxx
xx patternA xx
patternB
xxxxx
xx patternA xx
xxxxx
patternB
xxxxx

The result file should like like :
xxxxx
xx patternA xxpatternB
xxxxx
xx patternA xxxxxxxpatternB
xxxxx

# perl -pe 'BEGIN{$/="xxxxx\n"}chomp;s/\n/ /g;$_.=$/' xx.txt
xxxxx
xx patternA xx patternB xxxxx
xx patternA xx xxxxx
patternB xxxxx