CODEWITHSUNDEEP
linuxbashawksed
Austin
Austin

Reputation: 1255

Write blocks in a text file to multiple new files

I am trying to extract blocks in a text file and put them to new individual files. For example, consider the following file:

some junk lines

ABC: this is abc text
abc block text1
abc block text2
abc block text3

I dont care about this line

Text at start of block. I dont want this line also. 
ABC: this is another abc text
abc block text5
abc block text2
abc block text3
abc block text1

some other dont care line

I am interested in 'ABC' blocks. Every block has "ABC:" at beginning and new line at the end. So, I want to generate abc1.txt that contains:

ABC: this is abc text
abc block text1
abc block text2
abc block text3

and abc2.txt that contains:

ABC: this is another abc text
abc block text5
abc block text2
abc block text3
abc block text1

I tried using awk to get the blocks but having hard time in matching ending new line.

One option is to write a script that loops through each and every line in the file. I believe there is a better solution. Can someone please help? Thanks in advance!

Upvotes: 0

Views: 685

Answers (3)

Ed Morton
Ed Morton

Reputation: 204446

Your problem of blocks of text separated by blank lines is exactly what awks "paragraph mode" exists to handle and is activated by setting RS to the null string:

awk -v RS= '/^ABC:/{print > ("abc"++c".txt")}' file

The above will work if you don't have a lot of output files or if you're using GNU awk since it handles closing files for you when necessary. If you do have a lot of output files but can't get GNU awk then you just need to tweak it to:

awk -v RS= '/^ABC:/{close(f); f="abc"++c".txt"; print > f}' file

Upvotes: 1

Kent
Kent

Reputation: 195229

This one-liner should do the job:

awk '/^ABC/{p=1;close(fn);fn="abc"++i}!NF{p=0}p{print > fn}' file

With your example as input:

kent$  awk '/^ABC/{p=1;close(fn);fn="abc"++i}!NF{p=0}p{print > fn}' f

kent$  head abc*
==> abc1 <==
ABC: this is abc text
abc block text1
abc block text2
abc block text3

==> abc2 <==
ABC: this is another abc text
abc block text5
abc block text2
abc block text3
abc block text1

Note:

  • the close(fn) is necessary, if you have many "ABC" blocks, otherwise you got error msgs like "too many opened files"

Upvotes: 4

Raman Sailopal
Raman Sailopal

Reputation: 12887

awk '/^ABC:/,/^$/' filename

Search for all lines starting with ABC: (^ for the start of the line) to any thing with a blank line (^$) Use the , to lines to and from.

Upvotes: -2

Related Questions