I need help with a script that does the following operation. A text file contains the following information. AAA111 a line of some text another line of some text key: value1 END AAA222 a line of some text another line of some text key: value2 END AAA333 a line of some text another line of some text key: value3 END I want to extract multiple instances of a block of text (shown below) based on a particular key. AAA222 a line of some text another line of some text key: value2 END Let us say, there are 10 blocks with value2. These 10 blocks are to be moved to a different file from the original file.

If your blocks of texts were separated by blank (null) lines, then the following would work: sed -e '/./{H;$!d;}' -e 'x;/value2/!d;' myfile > outfile But in your example, I see that the second and third groups do not have a blank line between them. Soooooo ...... On the assumption that you are not interested in blank lines except to delimit your groups, I would first pass the file through sed to ensure that the END line is followed by a blank line, and to clean up any extraneous blank lines. Then the second sed can pick out the groups you want: sed -e /^$/d -e /^END$/G myfile | sed -e '/./{H;$!d;}' -e 'x;/value2/!d;' > outfile

Hi James, The script does exactly what I was looking for. Your help is very much appreciated. Joseph