Text processing

`column`

Format input into aligned columns or a table; turns CSV and whitespace data into readable output.

column [OPTION...] [FILE...]

10 of 10 examples run in Debian 13.7 on 2026-10-31 · bsdextrautils 2.41.5-0+deb13u1 · manual: column(1)

column has two modes. With -t it builds a table: it splits each line into fields (on whitespace, or on the characters given with -s) and pads every field so the columns line up. Without -t it fills a list of short items into as many columns as fit the screen width, like ls does.

column -t -s, is the fastest way to read a CSV in the terminal. The Linux version (from util-linux, in Debian's bsdextrautils package) adds much more: custom output separators, right-aligned columns, header names, hiding columns and even JSON output. It is not a CSV parser, though: it does not understand quoted fields with commas inside.

Examples

Read a CSV as an aligned table

$ column -t -s , data/users.csv | head -n 5

explain

output

id  name           email               dept         joined
1   Alice Moreau   alice@example.com   engineering  2021-04-12
2   Bob Okafor     bob@example.com     sales        2019-11-03
3   Chen Wei       chen@example.com    engineering  2023-01-20
4   Dana Kowalski  dana@example.com    support      2022-07-08

-t makes a table and -s , splits fields on commas. Each column is padded to its widest value, so the names and emails line up.

Align whitespace-separated data

$ column -t data/hosts.txt

explain

output

web01    10.0.0.11
web02    10.0.0.12
db01     10.0.0.21
cache01  10.0.0.31

Without -s, fields are split on runs of spaces and tabs, so ragged two-column data gets a straight second column.

Tidy up a command's output

$ printf '%s %s %s\n' NAME SIZE TYPE app.py 17 python helpers.py 3 python main.c 6 c | column -t

explain

output

NAME        SIZE  TYPE
app.py      17    python
helpers.py  3     python
main.c      6     c

Any command that prints space-separated fields can be piped through column -t for a neat table.

Draw visible column separators

$ column -t -s , -o ' | ' data/sales.csv | head -n 4

explain

output

date       | region | product | units | price
2026-03-01 | north  | widget  | 12    | 4.50
2026-03-01 | south  | gadget  | 3     | 19.99
2026-03-02 | east   | widget  | 7     | 4.50

-o ' | ' sets what goes between output columns (two spaces by default), giving a Markdown-like table.

Right-align numeric columns

$ column -t -s $'\t' -R 2,3 data/inventory.tsv

explain

output

name    qty  price
widget   12   4.50
gadget    3  19.99
gizmo     5  12.00

-s $'\t' splits on tabs (bash turns $'\t' into a real tab) and -R 2,3 right-aligns the quantity and price columns so the digits line up.

Name and hide columns

$ tail -n +2 data/users.csv | column -t -s , -N id,name,email,dept,joined -H email,joined | head -n 4

explain

output

id  name           dept
1   Alice Moreau   engineering
2   Bob Okafor     sales
3   Chen Wei       engineering

-N names the columns and prints the names as a header line; -H then hides columns by name. The file's own header is skipped with tail -n +2, otherwise it would appear twice.

Convert a table to JSON

$ tail -n +2 data/users.csv | head -n 2 | column -J -s , -n users -N id,name,email,dept,joined

explain

output

{
   "users": [
      {
         "id": "1",
         "name": "Alice Moreau",
         "email": "alice@example.com",
         "dept": "engineering",
         "joined": "2021-04-12"
      },{
         "id": "2",
         "name": "Bob Okafor",
         "email": "bob@example.com",
         "dept": "sales",
         "joined": "2019-11-03"
      }
   ]
}

-J writes JSON: an object whose key is the table name from -n, holding one object per row keyed by the column names from -N. All values are strings.

Fill a list into columns

$ seq 1 12 | column -c 40

explain

output

1	4	7	10
2	5	8	11
3	6	9	12

Without -t, items are arranged in columns down the page, as many as fit into the width given by -c (the terminal width by default). The gaps are tab characters.

Fill rows first

$ seq 1 12 | column -x -c 40

explain

output

1	2	3	4	5
6	7	8	9	10
11	12

-x fills across each row before starting the next, so the numbers read left to right.

Empty fields are kept

$ printf 'host,port,user\nweb01,,admin\ndb01,5432,\n' | column -t -s ,

explain

output

host   port  user
web01        admin
db01   5432

The missing port of web01 and the missing user of db01 stay as blank cells, so the columns don't shift. The BSD/macOS column merges adjacent separators instead.

Common pitfalls

column -t -s, is not a CSV parser: a quoted field such as "Smith, Jane" is split at its comma. Use a real CSV tool (python's csv module, mlr, csvlook) for such files.
BSD/macOS column treats several adjacent separators as one, so empty CSV fields shift the following values one column left. util-linux column on Linux keeps empty fields.
column reads all input before printing anything, because it needs every row to compute widths. It is useless with tail -f and slow on huge files.

GNU vs BSD and macOS

The examples above ran the GNU/Linux version. These are the differences that matter on macOS and the BSDs (written from their documentation, not run here).

The Linux (util-linux) column options -o, -R, -N, -H, -J, -n, -W and long options are not available in the BSD/macOS version, which has essentially -t, -s, -c and -x.

Options worth knowing

A selection, described in our own words. For the full list, see column(1) or column --help.

Option	What it does
`-t, --table`	Build a table: split lines into fields and align them.
`-s, --separator CHARS`	Characters that separate input fields in table mode (whitespace by default).
`-o, --output-separator STRING`	String placed between output columns (two spaces by default).
`-R, --table-right COLUMNS`	Right-align these columns (numbers or names).
`-N, --table-columns NAMES`	Comma-separated names for the columns.
`-H, --table-hide COLUMNS`	Do not print these columns.
`-O, --table-order COLUMNS`	Print columns in this order.
`-J, --json`	Output JSON; needs column names from -N.
`-n, --table-name NAME`	Name of the table, used as the JSON key.
`-W, --table-wrap COLUMNS`	Allow these columns to wrap their text.
`-c, --output-width WIDTH`	Output width in characters for fill mode.
`-x, --fillrows`	Fill rows before columns.
`-L, --keep-empty-lines`	Keep empty input lines instead of skipping them.

Examples

Common pitfalls

GNU vs BSD and macOS

Options worth knowing

Related