vi/comp.editors/blankline

From akf@awful.august.com (Andrew Fullford)
Subject: Re: Removing blank lines in vi
Date: Mon, 31 May 1993 04:25:36 GMT

>>>does anyone know how to remove blank lines from a file using vi?.

>Using
>	:g/^$/d

>should do the job.

This is splitting hairs, I know, but the original poster said "blank"
lines:

	:g/^[ ^I]*$/d

(that's "space tab").


From hansm@wsinti06.info.win.tue.nl (Hans Mulder)
Subject: Re: ddn.n.n.n.n. (oops, one too many) u (was: Removing blank lines in vi)
Date: 1 Jun 1993 21:21:27 +0200

In <1ufv9eINNb1o@uwm.edu> markh@csd4.csd.uwm.edu (Mark) writes:
>Maybe I wasn't totaly clear, in which case I'm sorry.  The question is, how
>do you automate this in vi: n.n.n.n.n. (until n fails)?

If you like to live dangerously, type:

:map q n.qm

This makes typing q do what you asked (n.n.n.n.n. until n fails).
And yes, typing u afterwards will ondo the last one.

I usually :w before I do this, and :unmap q when I'm done, just in case.

You might like to know about :set nowrapscan, which stops searches at
both ends of the buffer, which can be useful if you do things like this.

The sole purpose of the m at the end of the :map is to stop vi from
saying "No tail recursion".  I don't understand why tail recursion
is not allowed, while general recursion is.

HansM


From dattier@genesis.MCS.COM (DWT)
Subject: Re: Removing blank lines in vi
Date: 27 May 1993 14:14:18 -0500
Reply-To: dattier@genesis.mcs.com (DWT)

dberg@informix.com (David I. Berg) wrote in <dberg.738516173@puma> as others
have written in other articles:

|     :g/^$/d 
| 	
| will do the trick.

:v/./d   will do it with fewer characters and less shifting.

David W. Tamkin   Box 59297   Northtown Station, Illinois  60659-0297
dattier@genesis.mcs.com   CompuServe: 73720,1570   MCI Mail: 426-1818


From lau@auriga.rose.brandeis.edu (frankie t. k. lau)
Subject: Re: Removing blank lines in vi  < the surest way >
Date: Thu, 27 May 1993 20:14:25 GMT

tgcpwd@rwc.urc.tue.nl (Wim van Dorst/Prof. Penninger) writes:

>In article <1993May27.090951.65821@qut.edu.au> meilak@qut.edu.au writes:
>>does anyone know how to remove blank lines from a file using vi?.
>>locating empty lines can be done with 
>>	:s/^$/
>>the hard part is removing them!

>	:%g/^$/d


	From my belived book Unix in a Nutshell by O'Reilly.

	:g/^[  ]*$/d
             ^^
             /\
       space    tab

	this way, you can get rid of blank lines with or
without hidden space(s) and tab(s).

-Frankie


From gibson@netcom.com (Bob Gibson)
Subject: Re: multiple blank lines -> one blank line
Date: Thu, 15 Jul 1993 18:53:48 GMT

cat -S filename >newfilename
-- 
#include <disclaim.std>     /* I admit nothing! */
Bob Gibson -- gibson@netcom.com


From dattier@genesis.MCS.COM (David W. Tamkin)
Subject: Re: multiple blank lines -> one blank line
Date: 15 Jul 1993 14:24:29 -0500
Reply-To: dattier@genesis.mcs.com (DWT)

nancym@u.washington.edu (Nancy McGough) wrote in
<223t68$hpf@news.u.washington.edu>:

| I know I've seen this discussed before but I can't
| find the old messages.  I'd like to know how to
| convert multiple blank lines to one blank line
| using sed and using vi.  Also, is there a unix
| command like grep with a flag that will do this?

If you have Berkeley cat, cat -s will do it.  I've read that less -s does
the same thing but I've never seen a version of less where it worked.

NOTE that in SysV cat, the -s option means something else entirely.

sed '/./,/^$/!d' will do it, but if you're using a csh-based shell,
be sure to escape the exclamation point.  That has different results
from BSD cat -s if there are blank lines at the top; I'll go into further
detail on that if anyone is interested.

vi cannot do it with native vi or ex commands as far as I know because it
requires examining more than one line of text at a time, but it can with a
shell escape.

David W. Tamkin   Box 59297   Northtown Station, Illinois  60659-0297
dattier@genesis.mcs.com   CompuServe: 73720,1570   MCI Mail: 426-1818


From hansm@wsinti06.info.win.tue.nl (Hans Mulder)
Subject: Re: multiple blank lines -> one blank line
Date: 16 Jul 1993 19:39:20 +0200

In <224atd$s09@genesis.MCS.COM> dattier@genesis.MCS.COM (David W. Tamkin) writes:
>nancym@u.washington.edu (Nancy McGough) wrote in
><223t68$hpf@news.u.washington.edu>:

>| I know I've seen this discussed before but I can't
>| find the old messages.  I'd like to know how to
>| convert multiple blank lines to one blank line
>| using sed and using vi.  Also, is there a unix
>| command like grep with a flag that will do this?

>If you have Berkeley cat, cat -s will do it.
>NOTE that in SysV cat, the -s option means something else entirely.

>vi cannot do it with native vi or ex commands as far as I know because it
>requires examining more than one line of text at a time, but it can with a
>shell escape.

There is a native ex mode command to do it:

:g/^$/.,/./-j

, or, if you regard lines with only spaces and tabs as blank:

:g/^[ 	]*$/.,/[^ 	]/-j

The ^[ is really a ^ and a [, not an escape; between the [] are a
space and a tab.

HansM


From dattier@genesis.MCS.COM (David W. Tamkin)
Subject: Re: multiple blank lines -> one blank line
Date: 16 Jul 1993 13:40:56 -0500
Reply-To: dattier@genesis.mcs.com (DWT)

hansm@wsinti06.info.win.tue.nl (Hans Mulder) wrote in
<226p48$al7@wsinti06.info.win.tue.nl>:

| There is a native ex mode command to do it:
| 
| :g/^$/.,/./-j
| 
| , or, if you regard lines with only spaces and tabs as blank:
| 
| :g/^[ 	]*$/.,/[^ 	]/-j
| 
| The ^[ is really a ^ and a [, not an escape; between the [] are a
| space and a tab.

Thanks for the suggestion, Hans.  It works well except at the bottom of the
document.  Then, with wrapscan on, the ".,/./" search causes an "Addr 1 >
Addr 2" error; with wrapscan off, it causes a "no match to bottom" error. 
Either way, extra blank lines at the foot remain unchanged by the command;
one can remove them by hand.

It's a good alternative in a situation where you cannot (or don't want to)
escape to a shell.

David W. Tamkin   Box 59297   Northtown Station, Illinois  60659-0297
dattier@genesis.mcs.com   CompuServe: 73720,157From david@cats.ucsc.edu (David Michael Wright)
From: david@cats.ucsc.edu (David Michael Wright)
Newsgroups: comp.editors
Subject: Removing Blank lines
Date: 19 Dec 1993 09:20:18 GMT
Organization: University of California; Santa Cruz
Lines: 23



 have a quote from David Tamkin that explains how to remove blank lines:
sed '/./,/^$/!d'

If the lines are not truly empty but contain some spaces or tabs, cat -s
won't help either.  This will, though:

sed 's/[ 	]*$//
     /./,/^$/!d'

where the brackets enclose a space and a tab.

I am not sure, however, to apply this to my file. I tried !}sed ...
and %sed -c and so on, but it did not work. (Am running on System V
unix)
-- 
"There is nothing in the marginal conditions that     
    distinguish a mountain from a mole hill"            
                      Kenneth Boulding                             

					All comments  are mine---(David Wright) 
						david@cats.ucsc.edu.

From dattier@genesis.Mcs.Com (David W. Tamkin)
From: dattier@genesis.Mcs.Com (David W. Tamkin)
Newsgroups: comp.editors
Subject: Re: Removing Blank lines
Date: 19 Dec 1993 03:31:41 -0600
Organization: Contributor Account on MCSNet, Chicago, Illinois  60657
Lines: 39

david@cats.ucsc.edu (David Michael Wright) wrote in
<2f16ci$4kv@darkstar.ucsc.edu>:

|  have a quote from David Tamkin that explains how to remove blank lines:
| sed '/./,/^$/!d'

Actually, that was to squeeze successive multiple empty lines into one at
a time.

| If the lines are not truly empty but contain some spaces or tabs, cat -s
| won't help either.  This will, though:
| 
| sed 's/[ 	]*$//
|      /./,/^$/!d'
| 
| where the brackets enclose a space and a tab.

True.

| I am not sure, however, to apply this to my file. I tried !}sed ...
| and %sed -c and so on, but it did not work. (Am running on System V
| unix)

You mean within vi if you don't have BSD cat or GNU cat?

You cannot embed newlines, no matter what, in a shell command at the colon
prompt in vi or ex.  But there are at least two ways around it:

First method: put the sed commands into a sedfile; say you've named it
$HOME/sedfiles/squeeze.  Then 

:%!sed -f $HOME/sedfiles/squeeze

Second method: use sed's -e option to string the commands, like this:

:%!sed -e 's/[ 	]*$//' -e '/./,/^$/!d'

David W. Tamkin  P. O. Box 3284  Skokie, Illinois  60076-6284
dattier@mcs.com   CompuServe: 73720,1570   MCI Mail: 426-1818