UNIX for Dummies Questions & Answers

How would I split a file based on the location of a string, basically I want all entries above the string unix in this example

1
2
3
4
unix
5
6
7

Thanks,
Chuck

One way....I was nice and didn't put it in my standard csh...

#!/bin/ksh
file="file1"
while read -r xx
do
if [ "$xx" = "unix" ]
then
file="file2";
else
echo "$xx" >> $file ;
fi
done < $1

Original file has what you mentioned.

File1 ends up with
1
2
3
4
File2 ends up with
5
6
7
8

If you want the 'unix' in one of the files, just move the echo or add another.

I finally got this to work with just sed...

Code:

#! /usr/bin/sed -nf
s/unix/unix/
t loop
w file1
d
: loop
n
w file2
b loop

The script does work and so I thought that I might as well post it. I always have a tough time with these sed scripts though. I kinda fiddle with them until they work. It is very fast...it would take a custom c program to beat it. But it's hard to understand and maintain. So I actually prefer thehoghunter's solution.

To put the unix line in the first file, move the "w file1" up to be the first line in the script. To put that line in the second file, reverse the "n" and "w file2" lines.

This awk exits when it sees the unix line. To include the unix line into the file, put the print command first. If you want the unix grep string anchored to beginning-of-line, use /^unix/.

Code:

awk '/unix/{exit};{print}' myfile > file1