Global options

The line immediately following the .TS macro may contain any of the following global options (ignoring the case of characters – Unix tbl only accepts options with all characters lowercase or all characters uppercase), separated by spaces, tabs, or commas:

allbox

Enclose each item of the table in a box.

box

Enclose the table in a box.

center

Center the table (default is left-justified). The alternative keyword name centre is also recognized (this is a GNU tbl extension).

decimalpoint(c)

Set the character to be recognized as the decimal point in numeric columns (GNU tbl only).

delim(xy)

Use x and y as start and end delimiters for eqn(1).

doublebox

Enclose the table in a double box.

doubleframe

Same as doublebox (GNU tbl only).

expand

Make the table as wide as the current line length (providing a column separation factor). Ignored if one or more ‘x’ column specifiers are used (see below).

In case the sum of the column widths is larger than the current line length, the column separation factor is set to zero; such tables extend into the right margin, and there is no column separation at all.

frame

Same as box (GNU tbl only).

linesize(n)

Set lines or rules (e.g. from box) in n-point type.

nokeep

Don't use diversions to prevent page breaks (GNU tbl only). Normally tbl attempts to prevent undesirable breaks in boxed tables by using diversions. This can sometimes interact badly with macro packages' own use of diversions—when footnotes, for example, are used.

nospaces

Ignore leading and trailing spaces in data items (GNU tbl only).

nowarn

Turn off warnings related to tables exceeding the current line width (GNU tbl only).

tab(x)

Use the character x instead of a tab to separate items in a line of input data.

The global options must end with a semicolon. There might be whitespace between an option and its argument in parentheses.

Table format specification

After global options come lines describing the format of each line of the table. Each such format line describes one line of the table itself, except that the last format line (which you must end with a period) describes all remaining lines of the table. A single-key character describes each column of each line of the table. Key characters can be separated by spaces or tabs. You may run format specifications for multiple lines together on the same line by separating them with commas.

You may follow each key character with specifiers that determine the font and point size of the corresponding item, that determine column width, inter-column spacing, etc.

The longest format line defines the number of columns in the table; missing format descriptors at the end of format lines are assumed to be L. Extra columns in the data (which have no corresponding format entry) are ignored.

The available key characters are:

a,A

Center longest line in this column and then left-justifies all other lines in this column with respect to that centered line. The idea is to use such alphabetic subcolumns (hence the name of the key character) in combination with  L; they are called subcolumns because A items are indented by 1n relative to L entries. Example:

Result:

c,C

Center item within the column.

l,L

Left-justify item within the column.

n,N

Numerically justify item in the column: Units positions of numbers are aligned vertically. If there is one or more dots adjacent to a digit, use the rightmost one for vertical alignment. If there is no dot, use the rightmost digit for vertical alignment; otherwise, center the item within the column. Alignment can be forced to a certain position using ‘\&’; if there is one or more instances of this special (non-printing) character present within the data, use the leftmost one for alignment. Example:

Result:

If numerical entries are combined with L or R entries – this can happen if the table format is changed with .T& – center the widest number (of the data entered under the N specifier regime) relative to the widest L or R entry, preserving the alignment of all numerical entries. Contrary to A type entries, there is no extra indentation.

Using equations (to be processed with eqn) within columns which use the N specifier is problematic in most cases due to tbl's algorithm for finding the vertical alignment, as described above. Using the global delim option, however, it is possible to make tbl ignore the data within eqn delimiters for that purpose.

r,R

Right-justify item within the column.

s,S

Span previous item on the left into this column. Not allowed for the first column.

^

Span down entry from previous row in this column. Not allowed for the first row.

_,-

Replace this entry with a horizontal line. Note that ‘_’ and ‘-’ can be used for table fields only, not for column separator lines.

=

Replace this entry with a double horizontal line. Note that ‘=’ can be used for table fields only, not for column separator lines.

|

The corresponding column becomes a vertical rule (if two of these are adjacent, a double vertical rule).

A vertical bar to the left of the first key letter or to the right of the last one produces a line at the edge of the table.

To change the data format within a table, use the .T& command (at the start of a line). It is followed by format and data lines (but no global options) similar to the .TS request.

Column specifiers

Here are the specifiers that can appear in suffixes to column key letters (in any order):

b,B

Short form of fB (make affected entries bold).

d,D

Start an item that vertically spans rows, using the ‘^’ column specifier or ‘\^’ data item, at the bottom of its range rather than vertically centering it (GNU tbl only). Example:

Result:

e,E

Make equally-spaced columns. All columns marked with this specifier get the same width; this happens after the affected column widths have been computed (this means that the largest width value rules).

f,F

Either of these specifiers may be followed by a font name (either one or two characters long), font number (a single digit), or long name in parentheses (the last form is a GNU tbl extension). A one-letter font name must be separated by one or more blanks from whatever follows.

i,I

Short form of fI (make affected entries italic).

m,M

This is a GNU tbl extension. Either of these specifiers may be followed by a macro name (either one or two characters long), or long name in parentheses. A one-letter macro name must be separated by one or more blanks from whatever follows. The macro which name can be specified here must be defined before creating the table. It is called just before the table's cell text is output. As implemented currently, this macro is only called if block input is used, that is, text between ‘T{’ and ‘T}’. The macro should contain only simple troff requests to change the text block formatting, like text adjustment, hyphenation, size, or font. The macro is called after other cell modifications like b, f or v are output. Thus the macro can overwrite other modification specifiers.

p,P

Followed by a number, this does a point size change for the affected fields. If signed, the current point size is incremented or decremented (using a signed number instead of a signed digit is a GNU tbl extension). A point size specifier followed by a column separation number must be separated by one or more blanks.

t,T

Start an item vertically spanning rows at the top of its range rather than vertically centering it.

u,U

Move the corresponding column up one half-line.

v,V

Followed by a number, this indicates the vertical line spacing to be used in a multi-line table entry. If signed, the current vertical line spacing is incremented or decremented (using a signed number instead of a signed digit is a GNU tbl extension). A vertical line spacing specifier followed by a column separation number must be separated by one or more blanks. No effect if the corresponding table entry isn't a text block.

w,W

Minimum column width value. Must be followed either by a troff(1) width expression in parentheses or a unitless integer. If no unit is given, en units are used. Also used as the default line length for included text blocks. If used multiple times to specify the width for a particular column, the last entry takes effect.

x,X

An expanded column. After computing all column widths without an x specifier, use the remaining line width for this column. If there is more than one expanded column, distribute the remaining horizontal space evenly among the affected columns (this is a GNU extension). This feature has the same effect as specifying a minimum column width.

z,Z

Ignore the corresponding column for width-calculation purposes, this is, don't use the fields but only the specifiers of this column to compute its width.

A number suffix on a key character is interpreted as a column separation in en units (multiplied in proportion if the expand option is on – in case of overfull tables this might be zero). Default separation is 3n.

The column specifier x is mutually exclusive with e and w (but e is not mutually exclusive with w); if specified multiple times for a particular column, the last entry takes effect: x unsets both e and w, while either e or w overrides x.

Table data

The format lines are followed by lines containing the actual data for the table, followed finally by .TE. Within such data lines, items are normally separated by tab characters (or the character specified with the tab option). Long input lines can be broken across multiple lines if the last character on the line is ‘\’ (which vanishes after concatenation).

Note that tbl computes the column widths line by line, applying \w on each entry which isn't a text block. As a consequence, constructions like

.TS
c,l.
\s[20]MM
MMMM
.TE
    

fail; you must either say

.TS
cp20,lp20.
MM
MMMM
.TE
    

or

.TS
c,l.
\s[20]MM
\s[20]MMMM
.TE
    

A dot starting a line, followed by anything but a digit is handled as a troff command, passed through without changes. The table position is unchanged in this case.

If a data line consists of only ‘_’ or ‘=’, a single or double line, respectively, is drawn across the table at that point; if a single item in a data line consists of only ‘_’ or ‘=’, then that item is replaced by a single or double line, joining its neighbours. If a data item consists only of ‘\_’ or ‘\=’, a single or double line, respectively, is drawn across the field at that point which does not join its neighbours.

A data item consisting only of ‘\Rx’ (‘x’ any character) is replaced by repetitions of character ‘x’ as wide as the column (not joining its neighbours).

A data item consisting only of ‘\^’ indicates that the field immediately above spans downward over this row.

Text blocks

A text block can be used to enter data as a single entry which would be too long as a simple string between tabs. It is started with ‘T{’ and closed with ‘T}’. The former must end a line, and the latter must start a line, probably followed by other data columns (separated with tabs or the character given with the tab global option).

By default, the text block is formatted with the settings which were active before entering the table, possibly overridden by the m, v, and w tbl specifiers. For example, to make all text blocks ragged-right, insert .na right before the starting .TS (and .ad after the table).

If either ‘w’ or ‘x’ specifiers are not given for all columns of a text block span, the default length of the text block (to be more precise, the line length used to process the text block diversion) is computed as L×C/(N+1), where ‘L’ is the current line length, ‘C’ the number of columns spanned by the text block, and ‘N’ the total number of columns in the table. Note, however, that the actual diversion width as returned in register \n[dl] is used eventually as the text block width. If necessary, you can also control the text block width with a direct insertion of a .ll request right after ‘T{’.

Miscellaneous

The number register \n[TW] holds the table width; it can't be used within the table itself but is defined right before calling .TE so that this macro can make use of it.

tbl also defines a macro .T# which produces the bottom and side lines of a boxed table. While tbl does call this macro itself at the end of the table, it can be used by macro packages to create boxes for multi-page tables by calling it within the page footer. An example of this is shown by the -ms macros which provide this functionality if a table starts with .TS H instead of the standard call to the .TS macro.