Finding all information between keywords - linux

Question

I want to search a given word and retrieve all the surrounding lines between a pair of keywords:
I have the following data
NEW:
this is stackoverflow
this is a ghi/enlightening website
NEW:
put returns between paragraphs
indent code by 4 spaces
NEW:
here is this
most productive website
this is abc/enlightening/def
Now I want to retrieve all information between the two NEW which have the word "enlightening". That is, for the example input above I want the following output:
OUTPUT:
NEW:
this is stackoverflow
this is a ghi/enlightening website
NEW:
here is this
most productive website
this is abc/enlightening/def
I know that grep allows me to search a word-- but it retrieves only a specified number of lines e.g. 5 (specified by the user) above and below the given word. But how do I find out all the information between any keyword in linux("NEW" in this case). E.g. I specify here the delimiting keyword as "NEW" and call the information between any two new as paragraph. So, here my first paragraph is:
this is stackoverflow
this is a ghi/enlightening website
my second paragraph is:
put returns between paragraphs
indent code by 4 spaces
and so on.
Now I want all those paragraphs which have the keyword "enlightening" in them. i.e. I want the following output:
OUTPUT:
NEW:
this is stackoverflow
this is a ghi/enlightening website
NEW:
here is this
most productive website
this is abc/enlightening/def

Answer 1

The following AWK command should work (for mawk anyway -- POSIX doesn't seem to allow RS to be an arbitrary string):
awk -vRS='NEW:\n' -vORS= '/enlightening/ { print RS $0 }' data
Explanation:
-vFOO=BAR is a variable assignment.
Setting RS (Record Separator) to NEW:\n makes records be separated by NEW:\n instead of being lines.
Setting ORS to the empty string removes redundant blank lines after records on output. (Another option is to set it to NEW:\n, if having NEW:\n appear after the record is okay.)
/enlightening/ { print RS $0 } prints the record separator followed by the entire matching record ($0) for each record that contains "enlightening".
If having the separator appear after records is okay, then the command can be simplified to the following:
awk -vRS='NEW:\n' -vORS='NEW:\n' '/enlightening/' data
The default action when no action is specified is to print the record.
For strict POSIX compliance, appending lines to a temporary buffer while between two NEW:s and only printing that buffer if the search term was seen (could use a flag) should work, though it's more complicated.