manipulate \n

[BIS]

Hello,

Another silly question probably, but here goes.

Suppose I have a text file that looks like this:

;
42
;
33
;
437
;
16029
;
33

;
73
;
20
;
267
;
11029
;
45

and I want it to look like this:

;42;33;437;16029;33
;73;20;167;11029;45

how would I do that?

If I do cat file | tr -d '\n' it gives it all on one line.
There is a \n after every ; and number, with an additional \n between the sets. Any ideas?

 

[BIS]

..additionally...

i was hoping the tr -s command would do it, since thare are places where there are 2 \n's , but I can't get the syntax working.

 

[vgersh99]

nawk -f a.awk a.txt

#-------------------- a.awk
$0 ~ "^$" {print;next}
{ printf("%s", $0); }
END { print}
vlad
+---------------------------+
|#include<disclaimer.h> |
+---------------------------+

 

[BIS]

huh? will give it a go....

 

[ecasadella]

try:

#!/usr/bin/ksh
INFILE=$1
STRING=&quot;&quot;

cat $INFILE | while read LINE; do
echo $LINE | grep &quot;^$&quot;
if [ $? -eq 0 ]; then
STRING=&quot;${STRING}\n&quot;
else
STRING=&quot;${STRING}${LINE}&quot;
fi
done

echo $STRING


EDC

 

[BIS]

vlad's stuff does it perfectly. created a file called a.awk as specified and ran it - voila ! I'll be damned if I have any idea how though.
Guru vlad - thank you.

ecasadella: i'll try your version as well...

 

[DSMARWAY]

hi

or you can try this

cat file1 | while read aaa
do
print &quot;$aaa\c&quot;
if [[ $aaa = &quot;&quot; ]]
then
print &quot; \n&quot;
fi
done

 

[BIS]

vlad...?
would it be possible for you to quickly explain the above nawk programme?

 

[vgersh99]

sure - here it goes:

# nawk [NewAwk] invokation line
nawk -f a.awk a.txt

#-------------------- a.awk
# for any 'blank' line (a line [$0] that has ONLY a begining (^)
# and an end ($) with nothing else] print out a NEWLINE (print)
# and proceed to the next (next) record/line
$0 ~ &quot;^$&quot; {print;next}
#
# for any OTHER line [non-empty line] - print out the line
# with NO CR/LF (printf)
{ printf(&quot;%s&quot;, $0); }
#
# don't forget the LAST CR/LF for the LAST record/line
END { print}

#--------------------------------
I've posted a somewhat better/shorter/succinct solution in the other thread - that solution takes advantage of awk's RS [RecordSeparator] being an &quot;empty line&quot;.
vlad
+---------------------------+
|#include<disclaimer.h> |
+---------------------------+ vlad
+---------------------------+
|#include<disclaimer.h> |
+---------------------------+

 

[bigoldbulldog]

Here's a one-liner ( use backslash bang '\!' in some shells )

sed -e :a -e 'N' -e '$!{' -e '/\n$/!ba' -e '}' -e 's/\n//g' a.txt

Or as

sed '
:a
N
$!{
/\n$/!ba
}
s/\n//g
' a.txt Cheers,
ND

bigoldbulldog@hotmail.com

 

[toolkit]

So many methods. Here's my 2 cents:

perl -p -e 's/(.)\n/$1/;' test.txt > new.txt

Or an in place modification:

perl -i -p -e 's/(.)\n/$1/;' test.txt

Cheers, Neil