Usage of linux 'column' utility

I'm trying to generate an N columned table from a file of newline delimited data containing arbitrary text without any whitespace.

For example, I have a file of:

One
Two
Three
Four
Five
Six

I would like to display this data in the terminal like so:

One        Two
Three      Four
Five       Six

Unless I'm misunderstanding how column works, it doesn't seem to allow for anyway to specify the output table structure.

Without any arguments cat file | column simply shows them all on their own line just like cat does.

With the -t --table option, it shows them all on the same line cat file | column -t

I have tried every combination of arguments along with -t -C left -C right to try and get the output formatted into 2 columns.

I have also come very close to what I want with the pr tool: cat file | pr -t --columns 2

However, the formatting isn't quite right because some of the cells have whitespace that shifts the alignment off and pr doesn't really offer any other adjustments.

Can anyone offer a native shell/linux solution to take a list of text sequence and display them in a tabular format consisting of N columns?

Given:

$ cat file
One
Two
Three
Four
Five
Six

You can do:

$ paste -sd '\t\n' file | column -t
One    Two
Three  Four
Five   Six

(Alternatively, paste -sd '\t\n' - <file | column -t)

The key is that paste alternates between the two characters in the -d 'xy' string:

$ seq 10 | paste -sd '!|' -
1!2|3!4|5!6|7!8|9!10

So to get two columns with paste, alternate the column separator with the line separator:

$ paste -sd '\t\n' file
One	Two
Three	Four
Five	Six

If you want three columns, you would do:

$ paste -sd '\t\t\n' file
One	Two	Three
Four	Five	Six

An alternative to paste is this awk:

$ awk '{ getline line; printf "%s\t%s\n", $0, line }' file
# same output

The advantage to awk is the formatting between columns and line endings can be multiple characters.

Then to pad those two columns so they are straight, use column -t in a pipe after the awk or paste.

The column command was originally designed to fit the data to your terminal size. By default, it fills columns first, up to one less than the number of lines shown on your terminal screen, before moving to the next column. That's similar to the way text flows in a newspaper, for example. The few items in your example are not enough to fill the first column, which is why you don't discern the effect.

GNU column can also do the opposite by filling rows first, which you can request this with the -x option. You will definitely notice the difference here. But unless your screen is very narrow, you will also notice that the result is not really what you're looking for: column will decide the number of columns for itself (as it also did in the previous mode) based on the width of your terminal in characters and the widths of the various items.

Your best bet for doing this with column is to use its table mode, activated via option -t. Here you have to understand that column is just going to expect the input to already be formatted as a table -- one row per line, with columns separated by whitespace (by default). column has a large number of options for tweaking the result, but basically, it is just going to tidy that up so that it displays as neat columns.

To use column's table mode for your purpose, then, you first need to join adjacent rows of your input. For example:

sed 'N; s/\n/ /g' file | column -t

Of course, if your data contain whitespace then you'll need to choose a different field separator.

Beware that even in table mode, column is aimed at fitting the output to a fixed-width screen. You can control the width it uses via command-line option, and you'll likely need to do that if some of your items are lengthy or if you want more than a very few columns.

This is exactly what rs, the reshape command, excels at. Here, we just filter through rs 0 2 (unspecified rows of two columns):

#!/usr/bin/sh

printf '%s\n' One Two Three Four Five Six |
    rs 0 2

One    Two
Three  Four
Five   Six

It does the right thing when the entries vary in length, unlike simply formatting with tabs:

One           Two
Three         Four
Five          Six
Thirty-Three  Thirty-Four

also check this out :

paste - - < file | column -t

• paste - - < file merges every two lines of the file into a single line, separated by a tab!

output

One    Two
Three  Four
Five   Six

and if you want to specify a different number of columns,you can modify the paste accordingly,for example, to display the data in three columns, you can use:

paste - - - < file | column -t

or with awk like this

awk '{ printf "%-8s", $1 } NR % 2 == 0 { printf "\n" }' file 

and as you see you can edit the condition in the NR % 2 == 0 part accordingly,for example, to display the data in three columns, you can use NR % 3 == 0!

output

One    Two
Three  Four
Five   Six