Bash - listing files neatly - bash

Question

I have a non-determinate list of file names that I would like to output to the user in a script. I don't mind if it's a paragraph or in columns (like the out put of ls. How does ls manage it?). In fact I only have the following requirements:
file names need to stay on the same line (yes, that even means files with a space in their name. If someone is dumb enough to use a newline in a filename, though, they deserve what they get.)
If the output is formatted as a paragraph, I'd like to see it indented on the left and right to separate it from other text. Sort of like the way apt-get upgrade handles the list of packages to install.
I would love not to write my own function for this - at least not a complicated one. There are so many text formatting utilities in linux!
The utility should be available in the default Ubuntu install.
It should handle relatively large input, just in case. Something like 2000 characters or so?
It seems like a simple proposition, but I can't seem to get it to work. The column command is out simply because it can't handle large chunks of data. fmt and fold both don't care about delimiters. printf looks like it would work... if I wrote a script for it.
Is there a more flexible tool I've overlooked, or a simple way to do this?

Answer 1

Here I have a simple formatter that, it seems to me, is good enough
% ls | awk '
NR==1 {for(i=1;i<9;i++)printf "----+----%d", i; print ""
line=" " $0;l=2+length($0);next}
{if(l+1+length($0)>80){
print line; line = " " $0 ; l = 2+length($0) ; next}
{l=l+length($0)+1; line=line " " $0}}'
----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8
3inarow.py 5s.py a.csv a.not1.pdf a.pdf as de.1 asde.1 asdef.txt asde.py a.sh
a.tex auto a.wk bizarre.py board.py cc2012xyz2_5_5dp.csv cc2012xyz2_5_5dp.html
cc.py col.pdf col.sh col.sh~ col.tex com.py data data.a datazip datidisk
datizip.py dd.py doc1.pdf doc1.tex doc2 doc2.pdf doc2.tex doc3.pdf doc3.tex
e.awk Exit file file1 file2 geomedian.py group_by_1st group_by_1st.2
group_by_1st.mawk integers its.py join.py light.py listluatexfonts mask.py
mat.rix my_data nostop.py numerize.py pepp.py pepp.pyc pi.pdf pippo muore
pippo.py pi.py pi.tex pizza.py plocol.py points.csv points.py puddu puffo
%
I had to simulate input using ls because you didn't care to show how to access your list of files. The window width is arbitrary as well, but it's easy to provide a value to a -V width=... option of awk
Edit
I added a header line, an unrequested header line, to my awk script because I wanted to test the effectiveness of the (very simple) algorithm.
Addendum
I'd like to stress that the simple formatter above doesn't split "file names" like this across lines, as in the following example:
% ls -d1 b*
bia nconodi
bianconodi.pdf
bianconodi.ppt
bin
b.txt
% ls | awk '
NR==1 {for(i=1;i<9;i++)printf "----+----%d", i; print ""
line=" " $0;l=2+length($0);next}
{if(l+1+length($0)>80){
print line; line = " " $0 ; l = 2+length($0) ; next}
{l=l+length($0)+1; line=line " " $0}}'
----+----1----+----2----+----3----+----4----+----5----+----6----+----7----+----8
04_Tower.pdf 2plots.py 2.txt a.csv aiuole asdefff a.txt a.txt~ auto
bia nconodi bianconodi.pdf bianconodi.ppt bin Borsa Ferna.jpg b.txt
...
%
As you can see, in the first line there is enough space to print bia but not enough for the real filename bia nconodi, that hence is printed on the second line.
Addendum 2
This is the formatter the OP eventually went with:
local margin=4
local max=10
echo -e "$filenames" | awk -v width=$(tput cols) -v margin=$margin '
NR==1 {
for (i=0; i<margin; i++) {
line = line " "
}
line = line $0;
l = margin + margin + length($0);
next;
}
{
if (l+1+length($0) > width) {
print line;
line = ""
for (i=0; i<margin; i++) line=line " "
line = line $0 ;
l = margin + margin + length($0) ;
next;
}
{
l = l + length($0) + 1;
line = line " " $0;
}
}
END {
print line;
}'

Answer 2

Perhaps you're looking for /usr/bin/fold?
printf '%s ' * | fold -w 77 | sed -e 's/^/ /'
Replace the * with your list, of course; if your files are in an array (they should be; storing filenames in scalar variables is lossy), that'd be "${your_array[#]}".

Answer 3

If you have your filenames in a variable this will create 3 columns, you can change -3 to whatever number of columns you want
echo "$var" | pr -3 -t
or if you need to get them from the filesystem:
find . -printf "%f\n" 2>/dev/null | pr -3 -t
From what you stated in the comments, I think this may be what you are looking for. The find command prints the file or directory name along with a newline and you can put additional filtering of the filenames by piping through grep or sed prior to pr - the pr command is for print and the -3 states 3 columns and -t is for omit headers and trailers - you can adjust it to your preferences.