copying text between two unique text patterns



dear colleagues:
i have .rtf files of a collection of newspaper articles. each newspaper article starts with a variation of the phrase "document * of 20" and is separated from the next article with the character string "==================="

i would like to be able to take the text composing each news article from between these two patterns and dump them into separate, uniquely named files. i've been playing around with sed, grep, cut and csplit, but nothing seems to be working. i have the regular expressions developed to capture the two lines "document * of 20" and "--------" independently, but i can't figure out how to capture and play with the text between the two lines. i hope you can help.

yours,
simon j. kiss

hi simon,
though there could some other smarter solution,i have used the following approach to solve this problem.

assuming we have the contents of the file /tmp/mynewarticlefile.rtf as ,

cat /tmp/mynewarticlefile.rtf

html code:

times of india
edition-1
date:27 th may

document 1 of 20

all blah blah goes here
ad page
blah

================================

document 2 of 20

all blah blah goes here
ad page
blah

================================

document 3 of 20

all blah blah goes here
ad page
blah

================================
document 4 of 20

all blah blah goes here
ad page
blah

================================
end of the edition
thanks
editor

i have written the following script that process the above file to generate the output.
here the assumption is the document has 20 pages.

code:
#!/bin/ksh
let page=1
while [[ page -le 20 ]] ; do
sed -n /document\ $page/,/==========*/p /tmp/mynewarticlefile.rtf > /tmp/articlesplitpage-$page
((page=page+1))
done

upon execution of the above script i get 20 pages spilt according to the document no.

cat /tmp/articlespiltpage-1

html code:
document 1 of 20

all blah blah goes here
ad page
blah

================================

thanks,

hi.

for the sample data file "data1":

code:

document * of 20
hello

=====
document one of 20

world

=====
document 44 of 20

now is

=====
document "climatology review" in of 20

with no documents at the beginning of the time.

=====
i ran this script:

code:
#!/bin/sh

# @(#) s1       demonstrate csplit.

f=${1-data1}

csplit -k -s -z $f "/^document.*of/" {\*}

echo
for file in xx*
do
        echo
        echo "file: $file"
        head -3 $file |
        cat -n
done

exit 0
to produce this:

code:
% ./s1


file: xx00
     1  document * of 20
     2  hello
     3

file: xx01
     1  document one of 20
     2
     3  world

file: xx02
     1  document 44 of 20
     2
     3  now is

file: xx03
     1  document "climatology review" in of 20
     2
     3  with no documents at the beginning of the time.
this assumes that the lines "=====" are visual sugar ... cheers, drl