How to split a text file into 2 parts based on 3 (or more) newlines in bash

Question

I have a simple text file and I want to split it into 2 parts whenever I see 3 (or more) consecutive new lines. For example:

Fiona Gallagher
Frank Gallagher
Sheila Jackson



Dominique Winslow
Kermit

Should create 2 text files:

Fiona Gallagher
Frank Gallagher
Sheila Jackson

and

Dominique Winslow
Kermit

Though similar posts in SO address the problem of splitting text files into 2 parts (this PHP solution, this C# solution, and this Java solution) I could not find a solution in bash. It seems reasonable enough that a neat bash solution is out there somewhere (?). Thanks!

Ideally a mixture of grep, sed, cat and awk is best, but please post any other solution too. — OrenIshShalom, Sep 05 '18 at 06:29
That is not pure bash, so you might want to put specifically that in the question. — kabanus, Sep 05 '18 at 06:30
@kabanus could you re-post your csplit solution? I think there is a possible workaround with csplit. — OrenIshShalom, Sep 05 '18 at 07:17

kabanus · Accepted Answer · 2018-09-05T07:18:06.287

2

At the request of OP, a csplit solution that only works with 2 consecutive \n\n or more.

The easiest way would be to use the GNU csplit which any modern bash system has. Suppose bla contains:

Fiona Gallagher
Frank Gallagher
Sheila Jackson

Dominique Winslow
Kermit






a
v
b

then

csplit --suppress-matched -z bla "/^$/" '{*}'

will do what you want, creating

>cat xx00 
Fiona Gallagher
Frank Gallagher
Sheila Jackson
>cat xx01
Dominique Winslow
Kermit
>cat xx02
a
v
b

You can change the xx to any prefix using -f or --prefix=. The --suppress-matched makes sure the thing you split on (empty lines) will not be output and -z ensures consecutive empty lines do not form empty files. {*} Means split as many times as possible.

edited Sep 05 '18 at 07:18

answered Sep 05 '18 at 06:34

kabanus

24,623
6
41
74

can you enforce a minimum of 3 consecutive newlines? – OrenIshShalom Sep 05 '18 at 06:35
@OrenIshShalom Ah, no sorry. I'll remove the answer. I thought it said 2. – kabanus Sep 05 '18 at 06:36
Restored the answer as requested. – kabanus Sep 05 '18 at 08:11

How to split a text file into 2 parts based on 3 (or more) newlines in bash

1 Answers1