[pwharff]

Ok, here's a real stupid question:

How do I tell the "ls" command to output my data in 2 columns?

This is what I've already learned that will not work:

ls --width=45 (this works but is not reliable since the output size can change)
ls -1 (this outputs in single column format, ls -2 doesn't work (why!))
ls -x (this is ls across and I don't understand what it's useful for)

Someone please help me, I've spent over an hour and a half trying to get this to work! I know this is silly, but I need this to work and am willing to hear of suggestions.

[Wevah]

Shrink your terminal window?

No really...you might have to use a shell/Perl/whatever script.

[Partisan01]

Originally posted by pwharff:
Ok, here's a real stupid question:

How do I tell the "ls" command to output my data in 2 columns?

This is what I've already learned that will not work:

ls --width=45 (this works but is not reliable since the output size can change)
ls -1 (this outputs in single column format, ls -2 doesn't work (why!))
ls -x (this is ls across and I don't understand what it's useful for)

Someone please help me, I've spent over an hour and a half trying to get this to work! I know this is silly, but I need this to work and am willing to hear of suggestions.

If you want columns

ls -laF gives you a few columns with different items, you could always pipe it through grep and sed and spit out exactally what you want.

Nate

[Paul McCann]

Umm, maybe you *can't*. That is, it's probably not a built in. The available options look to be a single column (default when used into a pipe) or the multicolumn output that's used where the result is being displayed on a screen, or the "-x" option, which sorts across the page instead of down. Displaying in two columns isn't exactly a canonical sort of operation.

You can just pipe the output of ls into a script if you want to get the 2 column approach: here's the result of 2 minutes scripting. Ugly as hell, but does the job. Adjust as necessary, and pretty it up somebody: please.

Cheers,
Paul

% ls | perl -ne 'chomp($one=<>);chomp($two=<>);printf "%-30s %-30s\n",$one,$two'

[Rainy Day]

here's the result of 2 minutes scripting. Ugly as hell, but does the job. Adjust as necessary, and pretty it up somebody: please.

ls | perl -ne 'chomp($one=<>);chomp($two=<>);printf "%-30s %-30s\n",$one,$two'

Hmm... seems to drop the first file, and every third file thereafter. How about:

ls | perl -pe 's/(.*)\n/(pack "A30",$1).<>/e'

[Paul McCann]

Gack: thanks for the kick in the arse. What I vomited forth should have looked like

ls | perl -ne 'chomp;chomp($b=<>);printf "%-30s %-30s\n", $_,$b'

which has the virtue of actually printing all the files. But your pack version is definitely a bit cuter.

Cheers,
Paul

[VEGAN]

Originally posted by Paul McCann:
Gack: thanks for the kick in the arse. What I vomited forth should have looked like

ls | perl -ne 'chomp;chomp($b=<>);printf "%-30s %-30s\n", $_,$b'

which has the virtue of actually printing all the files. But your pack version is definitely a bit cuter.

Cheers,
Paul

You guys are making this all too complicated. All you need:

ls | rs 0 2

That's it!!!
`ls' output in 2 columns!!!

[Arkham_c]

Originally posted by VEGAN:
You guys are making this all too complicated. All you need:

ls | rs 0 2

That's it!!!
`ls' output in 2 columns!!!

The only problem is, if you have a space in a file name, this will not work.

Code:
[arkham@flybook ~/junk] touch "a b c d"
[arkham@flybook ~/junk] touch "d e f g"
[arkham@flybook ~/junk] ls | rs 0 2
a  b
c          d
d      e
f          g

So that's not quite it.

How about this:

Code:
[arkham@flybook ~/junk] ls -l
total 0
-rw-r--r--    1 arkham    staff           0 Sep 18 11:58 a b c d
-rw-r--r--    1 arkham    staff           0 Sep 18 11:58 d e f g
-rw-r--r--    1 arkham    staff           0 Sep 18 12:01 g h i j
-rw-r--r--    1 arkham    staff           0 Sep 18 12:01 j k l m
[arkham@flybook ~/junk] ls -1 -Q | sed 's/ /%%/g' | rs 0 2 | sed 's/%%/ /g'
"a b c d"  "d e f g"
"g h i j"      "j k l m"

[Paul McCann]

rs is a cool little utility that I knew nothing about: thanks for the pointer. While a bare "rs 0 2" won't do the trick if you use

ls | rs -c 0 2

the output looks OK to me. That causes the "columns" to be separated with a "^I" (whatever that is!) instead of space. I guess if your file name contains ^I you'd still be in trouble, but that doesn't sound terribly likely. If you know a character that's definitely not used in your filenames then inserting it directly after the -c will also work.

Cheers,
Paul

[Moonray]

'^I' or 'Control-I' is the TAB character. Although it's unlikely to have one in a file name, you can always be sure not to have a '/' in it. But then you can write as well ls -q | rs -e 0 2 (the -q to replace a newline (and other funny characters) with a '?').

However, the output looks bad if a filename is too long and then causes all lines to wrap. And that's the reason why ls has no -2 option: it either fits the output dynamically in as many columns (of different width) as possible to get you a short listing, or it serves you each entry on a single line if you find that easier to view or like to process it any further.

-

[Rainy Day]

Well if you just want a tab between filenames, then this will work:

ls | perl -pe 's/\n/"\t".<>/e'

Unlike rs, this works properly when filenames contain spaces.

[Moonray]

Originally posted by Rainy Day:
Well if you just want a tab between filenames, then this will work:

ls | perl -pe 's/\n/"\t".<>/e'

Unlike rs, this works properly when filenames contain spaces.

The above solutions do work properly when a filename contains spaces and they do not eat the final linefeed like yours does.

If someone really wanted tabs as column separator, rs would do that too with the whole command being shorter and using less resources.

-

[Rainy Day]

The above solutions do work properly when a filename contains spaces

And indeed it does appear to work! I stand corrected. I was thrown by man rs which states:

Code:
RS(1)                                                       RS(1)

NAME
       rs - reshape a data array

SYNOPSIS
       rs [ -[csCS][x][kKgGw][N]tTeEnyjhHm ] [ rows [ cols ] ]

DESCRIPTION
       Rs  reads  the standard input, interpreting each line as a
       row of blank-separated entries in an array, and writes it on the stan-
       dard output...

Looked to me, from that, that spaces would be treated as delimiters.

[Moonray]

I had to read the man page twice but it was worth it ... at the end they admit it has too many options

-