[SOLVED] Stylish text

K-Veikko · 08-06-2011, 05:27 PM

I have been playing around with sed to create human readable text files. Sometimes I encounter very long lines in textfiles.

How can I split a long line of text into several lines. – Just to make it more eye-pleasant.

Count 320 charcters from line beginning.
Find next .[space]
Replace [dot][space] with [dot][\n\n]
Go to #1 and continue counting from newly created [\n\n].

Prefererably a solution that suits in the pipe.

ta0kira · 08-06-2011, 05:45 PM

Maybe something like sed -r 's/(.{1,320}) (.)/\1\n\n\2/g', but that's for 1-320 characters per line.
Kevin Barry

edit: This might be what you're looking for: sed -r 's/(.{320}[^ ]*) (.)/\1\n\n\2/g'.

crts · 08-06-2011, 05:46 PM

Hi,

try this:

Code:

sed -r 's/(.{320}[^.]*\.)[[:blank:]]/\1\n\n/g' file

crts · 08-06-2011, 05:55 PM

Quote:

Originally Posted by ta0kira

Maybe something like sed -r 's/(.{1,320}) (.)/\1\n\n\2/g', but that's for 1-320 characters per line.
Kevin Barry

edit: This might be what you're looking for: sed -r 's/(.{320}[^ ]*) (.)/\1\n\n\2/g'.

Hi,

I think the OP wants to match a literal dot at the end followed by a space.

@OP:
Try also this alternatives:

Code:

sed -r 's/(.{320}[^.]*\.)[[:blank:]]+/\1\n\n/g' file
sed -r 's/(.{320}[^.]*\.)[[:blank:]]*/\1\n\n/g' file

I think you want the latter. It will split at the next dot even if it is not followed by a space.
The first one will only split if the dot is followed by at least one space or more. Both will remove all spaces following the dot.

K-Veikko · 08-25-2011, 05:23 PM

Quote:

Originally Posted by crts

Hi,I think the OP wants to match a literal dot at the end followed by a space.

This helped me.

Thank you very much for the quick answers. – Took quite a while to finally make my account active again. Sorry for the delay.

K-Veikko · 09-06-2011, 08:03 AM

Quote:

Originally Posted by K-Veikko

Code:

sed -r 's/(.{320}[^.]*\.)[[:blank:]]+/\1\n\n/g'

However. When struggling towards perfectness I met my limits again.

How can I make this linebreak happen if the script finds any of listed letters or any combination of them, actually any [^A-Za-z], followed by space. Especially

.[space] ".[space] ."[space]
?[space] "?[space] ?"[space]
![space] "![space] !"[space]
"[space]

crts · 09-06-2011, 09:00 AM

Quote:

Originally Posted by K-Veikko

However. When struggling towards perfectness I met my limits again.

How can I make this linebreak happen if the script finds any of listed letters or any combination of them, actually any [^A-Za-z], followed by space. Especially

.[space] ".[space] ."[space]
?[space] "?[space] ?"[space]
![space] "![space] !"[space]
"[space]

Do you still mean if it finds any of the characters after the first 320 characters?

Code:

sed -r 's/(.{320}[^[:punct:]]*[[:punct:]]+)[[:blank:]]+/\1\n\n/g'

K-Veikko · 09-06-2011, 12:37 PM

Quote:

Originally Posted by crts

Do you still mean if it finds any of the characters after the first 320 characters?

Code:

sed -r 's/(.{320}[^[:punct:]]*[[:punct:]]+)[[:blank:]]+/\1\n\n/g'

Any single character OR any combination of two (or more) characters FOLLOWED by space.

- I have not yet tested your solution.

crts · 09-06-2011, 01:51 PM

Quote:

Originally Posted by K-Veikko

Any single character OR any combination of two (or more) characters FOLLOWED by space.

- I have not yet tested your solution.

In this case you just might want to lose the first quantifier:

Code:

sed -r 's/([^[:punct:]]*[[:punct:]]+)[[:blank:]]+/\1\n\n/g'

The above will replace spaces that follow non-alphanumeric characters.