Remove new line starting with a numeric value and append it to the previous line

Shell Programming and Scripting
1

Hi,

i have a file with multiple entries. After some tests with sed i managed to get the file output as follows:

lsn=X-LINK-IN0,apc=661:0,state=avail,avail/links=1/1,
00,2110597,2094790,0,81,529,75649011,56435363,
lsn=TM1ITP1-AM1ITP1-LS,apc=500:0,state=avail,avail/links=1/1,
00,1019687575,1327955853,1,322,583,2810019516,1981897836,
lsn=TM1ITP1-AM1ITP2-LS,apc=501:0,state=avail,avail/links=1/1,
00,536022841,1334912507,0,277,966,2588691785,2512216939,
lsn=TM1ITP1-TMNLSTP1,apc=12309:1,state=avail,avail/links=4/4,
00,28422611,35850253,0,185,216,216892933,38975423,
01,66892965,50298069,0,250,8807,2591394100,803622868,
02,27488395,38094791,0,249,252,170335784,66760929,
03,27762094,19995809,0,251,258,216422353,2019547890,
lsn=TM1ITP1-TMNLSTP2,apc=12547:1,state=avail,avail/links=4/4,
00,80966204,16449140,0,47,121,1928336154,2818020786,
01,332655,171104,0,50,30,25432178,25498178,
02,55281799,11176343,0,93,145,4243472214,1907308254,
03,54060754,28585708,0,363,301,4174500371,4240529245,
lsn=X-LINK-IN1,apc=4882:1,state=avail,avail/links=1/1,
00,40085787,31652380,0,73,755,271023232,3802680330,

What i'm trying to perfrom now is to be able to shift any line starting with a number (in this case they all start with 0) to be shifted on to the next line. or maybe more simpler, all lines which do not begin with "lsn=" are to be appened to the previous line
Example:

lsn=X-LINK-IN0,apc=661:0,state=avail,avail/links=1/1,00,2110597,2094790,0,81,529,75649011,56435363,
....
lsn=TM1ITP1-TMNLSTP2,apc=12547:1,state=avail,avail/links=4/4,00,80966204,16449140,0,47,121,1928336154,2818020786,01,332655,171104,0,50,30,25432178,25498178,02,55281799,11176343,0,93,145,4243472214,1907308254,03,54060754,28585708,0,363,301,4174500371,4240529245,

Is there a way with sed or tr?

---------- Post updated at 02:47 PM ---------- Previous update was at 02:30 PM ----------

ok it seems this does the trick!

awk 'BEGIN {accum_line = "";} /^lsn=/{if(length(accum_line)){print accum_line; accum_line = "";}} {accum_line = accum_line " " $0;} END {if(length(accum_line)){print accum_line; }}'
2

Good you found a solution on your own. You might want to consider to put the preliminary sed stuff into that final awk script as well.

3

In the sample output you said you were trying to produce, there weren't any leading <space> characters at the start of your output lines and there weren't any <space> characters in the line where lines were joined. So, why does your script use:

{accum_line = accum_line " " $0;}

instead of just:

{accum_line = accum_line $0;}

?

4

Another idea

awk 'length{gsub(/\n/, ""); print RS $0 }' RS='lsn=' infile
5

This will work if you're using GNU awk .

But, according to the standards, the results are unspecified if the value assigned to RS is more than one character. On most BSD- and UNIX-systems, awk only looks at the first character of RS and ignores any remaining characters in that variable.

1 Like
7

Hi Don, you are correct.

in fact, I'm using the $0 now and it's working.

Thanks for the info.