Shell Programming and Scripting

Hi,

I'm beginner in UNIX
I would like to split file in separate files depending on Pattern.

Input file looks like:-

Code:

A
B
C
brfbeg
A
B
C
brfbeg
A
B
C
brfbeg
.
.

n so on

Expected output:-
File 1 -

Code:

A
B
C

File 2 -

Code:

A
B
C

n so on

I found $ awk '/START/{x="F"++i;}{print > x;}' INPUT
but it looks for pattern at the beginning & I don't have it in my file

Regards,
Rohit

Welcome Rohit_Mokal ,

Do you want to break to a new file each time you encounter the string brfbeg?

If that is correct, have a look at csplit

We have something that does this, so I have adjusted it to see if it helps:-

Code:

filecnt=`grep -c "^brfbeg" input_file`
((filecnt=$filecnt-1))
csplit -f my_filename -n 5 -s input_file "/^brfbeg/" {$filecnt}

This will create a set of files called my_filename.00000, my_filename.00001, my_filename.00002, my_filename.00003, etc.

The -n flag gives you the number of digits to use.

It will keep use the string to split on as the first line, but hopefully that's not too much of an issue.



Does this meet your needs?

Robin

Well, your attempt was not too far off. Try

Code:

awk '
NR==1           {x="F"++i}       
/brfbeg/        {x="F"++i
                 next 
                }
                {print > x}
' file

Depending on what OS you're using, what version of awk you're using, and the number of files to being created, this could run out of file descriptors. And if you're creating more than 9 files, you probably want to format your file numbers with leading zeros so all of the filenames will be the same length (and so file* will expand to a list of files with contents in the same order as the contents of the original file). Assuming you'd like 3 decimal digits in your output filenames, I'd suggest the slightly more complicated:

Code:

Code:

awk -v n=3 '
NR==1           {x=sprintf("File%0*d", n, ++i)}       
/brfbeg/        {close(x)
                 x=sprintf("File%0*d", n, ++i)
                 next
                }
                {print > x}
' file

If you want a different number of digits in your filenames, change the number shown in red to the number of digits you want.

And, as always, if someone wants to try this on a Solaris/SunOS system, change awk to /usr/xpg4/bin/awk.

Thanks everyone for your response!

I tried solution from Don & it worked fine. We r planning to split input file in 1000 small files & solution works perfect with n =4.

Hi,

Could you please let me know how to direct output files created in above awk (Post from Don) to different Directory?

You can add a path - either as a string constant or as a variable defined in the way n is defined - in front of the x file name, either in the redirected print command, or within the sprintf function parameters.