Shell Programming and Scripting

Although my sed skills are gradually developing, thanks in large part to this forum, I'm having a hard time dealing with pattern space and looping, which I suspect is what I'll need a better handle on to figure out my current issue, which is converting a multi line file like this:

12345678|ONE%TWO%THREE

into

12345678|ONE
12345678|TWO
12345678|THREE

The % delimited items to the right of the pipe is an unlimited range of elements. The left of the pipe will always contain an 8 digit number. My thought was to perhaps make two passes (maybe creating an intermediate file) that first just substituted the % delimiters with a pipe and carriage return, e.g.

Code:

 
s/%/\
|/g

Which would give me this...

12345678|ONE
|TWO
|THREE

Then on the next pass somehow grab just the number in the pattern space and prepend it to subsequent lines that do not begin with a number (the % delimited elements will always be alpha characters.)

And that's where I'm stuck. Any snippets or even just strategy suggestions would be greatly appreciated. Although this appears to be pretty low hanging fruit for the likes of perl, I'd like to hang with sed to the extent that it's possible... for the challenge if nothing else :-).

Thanks in advance,
Al

How about:

Code:

$ cat data
12345678|ONE%TWO%THREE
$ cat sedder
#! /usr/bin/ksh
exec < data
sed ':l;s/\([0-9][0-9]*\)\(.*\)\(%\)\(.*\)/\1\2\n\1|\4/;t l'
$
$ ./sedder
12345678|ONE
12345678|TWO
12345678|THREE
$

My sed lets me do \n for a newline so I went with that.

Another approach, but with awk:

Code:

awk -F\| '{gsub("%","\n"$1"|")}1' file

an awk version:

Code:

 
echo "12345678|ONE%TWO%THREE" | awk -F "[|%]" '{for(i=2;i<=NF;i++) { print $1,$i}}' OFS="|"

Thanks for the great responses guys - looking at various solutions is a great learning tool. I was further off from a solution than I thought :-)
Al