Shell Programming and Scripting

I have data of the following form. It starts its first line with a '{' and ends the last line with a '}'. What I need to do is to join every line starting with the '{' and ending with the '} into one line in the output file. If it was just this one block, I could do a 13 joins in vi, however it contains thousands of blocks of data. Any help would be greatly appreciated.

{F,10,A,R,E,0600,0400,"XXXX"|
T,01, 13,V, 0, 45,00,1,01,01,B,L,0,0 |
T,02, 14,V, 15, 45,00,1,01,01,B,L,0,0 |
T,03, 40,V, 59, 15,00,1,02,01,B,L,0,0 |
T,04, 30,V, 102, 15,00,1,01,01,B,L,0,0 |
T,05, 35,V, 128, 15,00,1,01,01,B,L,0,0 |
T,06, 10,V, 163, 50,00,1,01,01,B,L,0,0 |
T,07, 11,V, 203, 45,00,1,02,02,B,L,0,0 |
T,08, 4,V, 400, 175,00,1,01,01,B,L,0,0 |
T,09, 44,V, 420, 0,00,1,01,01,B,L,0,0 |
T,10, 31,V, 460, 15,00,1,02,01,B,L,0,0 |
T,11, 29,V, 485, 15,00,1,02,01,B,L,0,0 |
T,12, 36,V, 510, 15,00,1,02,01,B,L,0,0 |
T,13, 21,V, 550, 130,00,1,02,01,B,L,0,0 |
}

something to start with - assuming no embedded "blocks":

nawk -f sw.awk file

sw.awk:

Code:

substr($0,1,1) == "{" && substr($0,length($0),1) == "|" {
   start=1;
}
substr($0,1,1) == "}" && start{ print; start=0;next}
start {printf $0;next}
1

And here's a sed solution that should work with most newer seds....

Code:

#!/bin/sed -nf

H
/}/ {
  x
  s/\n//g
  p
}

Invoke with
./foo.sed my_file

Cheers
ZB

You solution worked great. Thanks very much.

We can do with awk as,

Code:

awk '/^{/,/^}/ { printf $0 }END { print "\n"}' <filename>

HTH.