%% LyX 2.4.2.1 created this file.  For more info, see https://www.lyx.org/.
%% Do not edit unless you really know what you are doing.
\documentclass[english,tableposition=top]{report}
\usepackage{lmodern}
\renewcommand{\sfdefault}{lmss}
\renewcommand{\ttdefault}{lmtt}
\usepackage[T1]{fontenc}
\usepackage{textcomp}
\usepackage[latin9]{inputenc}
\setcounter{secnumdepth}{3}
\setcounter{tocdepth}{3}
\usepackage{color}
\definecolor{shadecolor}{rgb}{0.667969, 1, 1}
\usepackage{babel}
\usepackage{array}
\usepackage{booktabs}
\usepackage{framed}
\usepackage{url}
\usepackage{varwidth}
\usepackage{amsmath}
\usepackage{makeidx}
\makeindex
\usepackage[pdfusetitle,
 bookmarks=true,bookmarksnumbered=true,bookmarksopen=true,bookmarksopenlevel=2,
 breaklinks=false,pdfborder={0 0 1},backref=section,colorlinks=true]
 {hyperref}

\makeatletter

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% LyX specific LaTeX commands.
%% Because html converters don't know tabularnewline
\providecommand{\tabularnewline}{\\}
%% Variable width box for table cells
\newenvironment{cellvarwidth}[1][t]
    {\begin{varwidth}[#1]{\linewidth}}
    {\@finalstrut\@arstrutbox\end{varwidth}}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Textclass specific LaTeX commands.
\newenvironment{centred}%
  {\begin{center}}{\end{center}}
\newenvironment{lyxcode}
	{\par\begin{list}{}{
		\setlength{\rightmargin}{\leftmargin}
		\setlength{\listparindent}{0pt}% needed for AMS classes
		\raggedright
		\setlength{\itemsep}{0pt}
		\setlength{\parsep}{0pt}
		\normalfont\ttfamily}%
	 \item[]}
	{\end{list}}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% User specified LaTeX commands.
\usepackage{numerica}
\usepackage{numerica-tables}
\usepackage{upquote}
\usepackage{xfrac}
\usepackage[indentunit=1em,totoc,columnsep=1.5em,
initsep=10.25pt plus 5pt minus 3pt]{idxlayout}
\usepackage{diagbox}
\reuse{}

\makeatother

\begin{document}
\title{\texttt{numerica-tables}~\\
{\large version 3.2.0}}
\author{Andrew Parsloe\\
(\url{ajparsloe@gmail.com})}
\maketitle
\begin{abstract}
The \texttt{numerica-tables} package enables the creation of multi-row,
multi-column tables of values of mathematical functions. Key\textendash value
assignments allow presentation in a wide variety of table styles,
both vertically by column or horizontally by row, within the `formal
table' framework of the \texttt{booktabs} package. \texttt{numerica-tables}
requires the prior loading of the \texttt{numerica} package.\medskip{}
\end{abstract}
\noindent{}%
\noindent\begin{minipage}[t]{1\columnwidth}%
\begin{shaded}%
\begin{itemize}
\item This document applies to version 3.2.0 of \texttt{numerica-tables}
and requires the \texttt{booktabs} package.
\item Version 3 of \texttt{numerica} needs to be loaded before \texttt{numerica-tables};
(\texttt{numerica} requires \texttt{amsmath} and \texttt{mathtools}).
\item I refer many times in this document to tables included in \emph{Handbook
of Mathematical Functions}, edited by Milton Abramowitz and Irene
A. Stegun, Dover, 1965, and referenced as \emph{HMF}.
\item Version 3.2.0 of \texttt{numerica-tables}
\begin{itemize}
\item adds a setting \verb`transpose` to convert a table with function
values in columns into a table with function values in rows;
\item adds the package options \verb`rules` and \verb`norules` to specify
the default pattern of horizontal rules;
\item adds new settings and enhances existing settings to adjust the appearance,
size and positioning of items in both the row variable column and
the header row; 
\item enhances and adjusts elements of fraction form output (e.g. with a
`semi-verbatim' setting);
\item fixes a bug when tabulating factorials;
\item updates documentation.
\end{itemize}
\item Version 3.1.0 of \texttt{numerica-tables}
\begin{itemize}
\item adds an index to the documentation;
\item adds the ability to round table entries to different values.
\end{itemize}
\item Version 3.0.0 of \texttt{numerica-tables}
\begin{itemize}
\item adds the ability to use as row variable values numbers or expressions
from a comma list, a macro, a file, or a step function, to be displayed
either as values or verbatim;
\item adds the ability to suppress the header row;
\item is compatible with the additional features of \texttt{numerica} version
3.0.0, including the decimal comma and fraction-form output.
\end{itemize}
\end{itemize}
\end{shaded}%
\end{minipage}

\tableofcontents{}

\chapter{Introduction}

\label{chap:Introduction}Entering 
\begin{verbatim}
    \usepackage[<options>]{numerica}
    \usepackage[<options>]{numerica-tables}
\end{verbatim}
\noindent in the preamble of a document gives access to a command
for creating tables of function values in a wide variety of styles.
Contrary to previous practice, from version 3.0.0 of \texttt{numerica-tables},
the \texttt{numerica} package is not loaded automatically but must
be loaded explicitly (as above), with options if desired, \emph{before}
\texttt{numerica-tables}. It is \emph{essential} that the version
of \texttt{numerica} loaded is version 3. 

All tables are `formal tables' in the sense of the \texttt{booktabs}\index{booktabs@\texttt{booktabs}}
package, which is loaded automatically. Such tables have no vertical
rules and few horizontal rules. Quoting from the documentation for
that package:\index{rules@\texttt{rules}!edicts}
\begin{enumerate}
\item Never, ever use vertical rules.
\item Never use double rules.
\end{enumerate}

\section{Package options}

\label{sec:Package-options}New \index{package options}\index{rules@\texttt{rules}!package option}\index{norules@\texttt{norules}!package option}
with version 3.2 of \texttt{numerica-tables} are the package options
\verb`rules` and \verb`norules`, and \verb`Q?*`, which is more
complicated and discussed later (\S\ref{sec:Star-option}). The \verb`rules`
option determines which (if any) horizontal rules are drawn in a table
if none are explicitly set for that table. Because most tables in
this document are set amongst text, they are delimited by top and
bottom rules, and one beneath the header row. For consistency with
earlier versions, `out of the box' the \verb`rules` package option
is initialized to \verb`rules=ThB`. 

For the user this may not be what is wanted. You may prefer by default
to insert a rule only beneath the header row of your tables in which
case the package option would be \verb`rules=h`. For other possible
values for the package option see the identical values of the \verb`rules`
\emph{setting} discussed in §\ref{subsec:Rules:-rules-setting}. If,
as a default, you want no rules at all in your tables, you can either
give \verb`rules` an empty value (\verb`rules= ` ) or use the \verb`norules`
package option:
\begin{verbatim}
    \usepackage[norules]{numerica-tables}
\end{verbatim}

\section{Table structure}

I\index{tables!structure} take as my source of models of mathematical
tables those presented in \emph{Handbook of Mathematical Functions},
edited by Milton Abramowitz and Irene A. Stegun, Dover, 1965, not
because the typesetting is elegant (it often is not) but because \emph{HMF}
displays a wide variety of table styles. The editors of that volume
were faced with a host of different problems requiring a host of different
solutions. The \verb`\nmcTabulate` command of \texttt{numerica-tables}
aims to reproduce most of those different solutions, within \texttt{booktabs}\index{booktabs@\texttt{booktabs}}
elegance.

Creating a table presumes we have a function or functions we wish
to tabulate. The values a function takes will generally depend on
a primary parameter and, possibly, a number of secondary parameters
(which is where much of the complexity comes from). Books of mathematical
tables are structured (nearly always) in \emph{columns} (but see below),
and we read (nearly always) \emph{down} a column as the primary parameter
is incremented, generally in regular steps. We need to decide on the
range of values the primary parameter will take and how fine-grained
the tabulation will be \textendash{} what the step size of its increments
will be. Assigning different values to a second parameter generates
a second, third, \ldots{} column of function values. Sometimes rather
than a second parameter, a second, third, \ldots{} function of the
first parameter is tabulated in the successive columns \textendash{}
like adjacent columns of different trigonometric functions. 

In this document the first parameter is called the \emph{row variable}\index{row variable}
\textendash{} its value determines which row we are in; the second
parameter, if present, is called the \emph{column variable\index{column variable}}
\emph{\textendash{} }its value determines which column we are in.
A table generally (but not always) presents the values of the row
variable in the first column, the \emph{row variable column}, sometimes
in distinctive type (e.g. bolded). The values of the column variable
are presented in a \emph{header row} above the table body of function
values. Above the header row there may be a \emph{title row} and perhaps
a \emph{subtitle row} where other explanatory material can be displayed.
Sometimes there is a \emph{footer row} beneath the table body. Vertical
rules are absent, horizontal rules used sparingly \textendash{} for
example, at the top and bottom of the table, or under the header row,
but not in the body of the table.

\subsection{Transposed tables}

From\index{tables!`horizontal'/`vertical'} version 3.2 of \texttt{numerica-tables}
it is also possible to present tables `horizontally', in which function
values are read in rows (generally) left-to-right across the page.
\texttt{numerica-tables} handles this situation by constructing the
table in the usual column-based `vertical' manner and then treating
the columns of calculated function values as rows; the row variable
column and the header row swap roles. The placement of the new row
variable column and header row and the `decoration' of the values
contained therein (font, alignment, etc.) is all done \emph{after}
transposition.

The process is effected through the setting \verb`transpose`\index{transpose@\texttt{transpose}}
and is discussed at §\ref{subsec:Transpose}, but examples are presented
throughout the document.

\subsection{\texttt{tabular} environment}

The tables created by \texttt{numerica-tables} use the \verb`tabular`\index{tabular environment@\texttt{tabular} environment}
environment of \LaTeX . This means that the vertical separation between
rows can be altered by means of the parameter \verb`\arraystretch`\index{arraystretch@\texttt{\textbackslash arraystretch}}.
Writing, for instance,
\begin{verbatim}
    \renewcommand\arraystretch{1.2}
\end{verbatim}
increases the separation between rows by a scale factor of $1.2$.
An example of use occurs in §\ref{subsec:rmath}.

The separation between columns is determined by the parameter \verb`\tabcolsep`\index{tabcolsep@\texttt{\textbackslash tabcolsep}}
which is half the width of the space between columns and defaults
to $6$ pt. To change \verb`\tabcolsep` you can use either of the
commands \verb`\addtolength` or \verb`\setlength`. Since the default
is $6$pt, the commands 
\begin{verbatim}
    \addtolength\tabcolsep{3pt}
    \setlength\tabcolsep{9pt}
\end{verbatim}
have the same effect and increase the half-width between columns to
$9$ pt; see the final example in §\ref{subsec:Nudgingtheheaders:-chnudge}.

After changing either of these parameters you will need to reset it
back to the default if you do not want all subsequent tables to be
stretched vertically or horizontally as the case may be.

\section{Shared syntax}

The \verb`\nmcTabulate`\index{nmcTabulate@\texttt{\textbackslash nmcTabulate}}
command, short-name form \verb`\tabulate`\index{tabulate@\texttt{\textbackslash tabulate}|see{\texttt{\textbackslash nmcTabulate}}},
shares the syntax of \verb`\nmcEvaluate` (see \texttt{numerica.pdf}).
When all options are used the command looks like 
\begin{verbatim}
    \nmcTabulate*[settings]{expr.}[vv-list][num. format]
\end{verbatim}
\selectlanguage{english}%
\begin{enumerate}
\item \verb`*` optional switch; if present ensures a single number output
with no formatting, or an appropriate error message if the single
number cannot be produced; see §\ref{sec:Star-option};
\item \verb`[settings]` \emph{mandatory }comma-separated list of \emph{key=value
}settings; this option is at the heart of creating a table of function
values, and is discussed in the next chapter;
\item \verb`{expr.}` mandatory argument specifying the mathematical expression
or expressions in \LaTeX{} form to be tabulated;
\item \verb`[vv-list]`\index{vv-list} \emph{mandatory} comma-separated
list (or semicolon-separated\index{decimal comma!semicolon use} list
if the \verb`comma` package option\index{comma package option@\texttt{comma} package option}
is used with \texttt{numerica}) of \emph{variable=value }items, in
particular containing the initial value of the row variable (essential)
and column variable (if one is used); 
\item \verb`[num. format]` optional format specification for presentation
of the numerical results (rounding, padding with zeros, scientific
notation, fraction-form output); see §\ref{subsec:Trailing-optional-argument}.\index{rounding value}
\end{enumerate}
Unlike \verb`\nmcEvaluate` (the main command in\texttt{ numerica}),
for \verb`\nmcTabulate` the two apparently optional arguments straddling
the main argument (\verb`settings` and \verb`vv-list`) are \emph{essential}.
Although both are delimited by square brackets, that is in order to
draw on the {\ttfamily\verb`numerica`} code. Each argument
contains items \emph{necessary} for the construction of any table
of function values.

Should \texttt{numerica} be loaded with the \verb`comma` package
option\index{comma package option@\texttt{comma} package option},
numbers in tables will be displayed with a decimal comma\index{decimal comma!semicolon use}.
In this case, to avoid confusion, items in the vv-list\index{vv-list!and decimal comma}
\emph{must} be separated by a semicolon. Similarly, $n$-ary functions
\textendash{} \verb`\max`, \verb`\min` and \verb`\gcd`\index{n-ary@$n$-ary functions}
\textendash{} must use semicolons as their argument separator. From
version 3.2 of \texttt{numerica-tables} row variable data in the form
of lists or files for the \verb`rdata` and \verb`rfile` settings
do not need to be \emph{comma} separated. If the \verb`comma` package
option\index{comma package option@\texttt{comma} package option}
is used, the semicolon is assumed (see §\ref{subsec:rdata-rfile-rverb})
or some other item separator can be set with the (new in version 3.2)
\verb`rdelim` setting.

Although math environments are significant for \verb`\nmcEvaluate`,
they should be avoided with \verb`\nmcTabulate`\index{nmcTabulate@\texttt{\textbackslash nmcTabulate}}\index{environments}.
Placing a \verb`\tabulate` command within a math environment, or
vice versa, is likely to cause a \LaTeX{} error. From version 3.2 of
\texttt{numerica-tables}, it is possible to specify the math style
(\verb`\displaystyle`, \verb`\textstyle` or \verb`\scriptstyle`)
of entries in the row variable column and in the header row. Function
values (numbers) in the table cells are presented between inline math
delimiters (\verb`$` \verb`$`).

\chapter{\texttt{\textbackslash nmcTabulate} settings}

Just as \verb`\nmcTabulate` shares the syntax of \verb`\nmcEvaluate`
(of \texttt{numerica}), it also shares the \emph{settings} of the
latter command \textendash{} although not all will be relevant. See
the associated document \texttt{numerica.pdf} for a list of those
settings and associated discussion. I will point out instances of
their use in the following examples. 

In addition to the shared settings, there are many specific to \verb`\nmcTabulate`,
which is what this document is about. They are discussed in groups
in later sections, some in more than one place. For the main discussion
of row variable settings, see immediately below; for column variable
settings see §\ref{sec:Column-variable-settings}; for multi-function
tables see §\ref{sec:Multiple-function-tables}; for whole-of-table
formatting see §\ref{sec:Whole-of-table-formatting}; for formatting
the function values in table cells see §\ref{sec:Function-value-formatting}.

\section{Row variable settings}

\label{sec:Row-variable-settings}Okay, assume we have chosen the
function we want tabulated. Now we need to choose the row variable,
how fine-grained we want the tabulation to be and between which values.

\subsection{Row variable specification: uniform case}

\label{subsec:Row-var-spec-uniform}\index{row variable!start, step, stop values|(}The
\emph{row} variable is set in the settings option of the \verb`\tabulate`
command with the key \verb`rvar`\index{rvar@\texttt{rvar}} \textendash{}
say \verb`rvar=x`. What value to start tabulating from is specified
in the vv-list\index{vv-list!row/col.variables@row/col.\ variables}\index{row variable!initial value in vv-list}
\textendash{} \verb`x=0` perhaps \textendash{} and so does not need
a specific key, but what value to tabulate to, \verb`rstop`\index{rstop@\texttt{rstop}},
and the step size, how fine-grained the tabulation is to be, \verb`rstep`\index{rstep@\texttt{rstep}},
do need to be specified in the settings option. In the uniform case
(which makes up the overwhelming majority of cases in \emph{HMF})
the step size is constant. It does not change as the value of the
row variable changes. (The non-uniform case, available from version
3.0.0 of \texttt{numerica-tables}, is discussed in §\ref{subsec:Row-var-spec-nonuniform}
below. Quite different keys are required.)

The two tables in the example below tabulate $\sin x$ and $\cos x$
between $0$ and $1$ in increments of $0.2$. By placing the start
value of the tabulation variable in the vv-list, the possibility is
opened for other parameters in more complicated functions to depend
on the row variable. Although it will often be the first entry in
the vv-list, it does not need to be. 

\begin{table}[t]
\centering
\centering{}\caption{Row variable specification}
\begin{center}
\begin{tabular}{ll>{\raggedright}p{4cm}>{\raggedright}p{3cm}}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small comment}\tabularnewline
\midrule
{\small\texttt{rvar}} & {\small token(s)} & {\small row variable} & \tabularnewline
{\small\texttt{rstep}} & {\small real num.} & {\small step size} & \tabularnewline
{\small\texttt{rstop}} & {\small real num.} & {\small stop value} & {\small excludes }{\small\texttt{rows}}\tabularnewline
{\small\texttt{rows}} & {\small int} & {\small number of rows} & {\small excludes }{\small\texttt{rstop}}\tabularnewline
{\small\texttt{rspec}} & {\small comma list} & {\small\texttt{\{rvar}}{\small , }{\small\texttt{rstep}}{\small , }{\small\texttt{rows\}}} & {\small short form spec.}\tabularnewline
{\small\texttt{rround}} & {\small int} & {\small rounding} & {\small default: }{\small\texttt{1}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}
The initial value of the row variable may depend on other quantities
which must necessarily precede it \textendash{} lie to the right of
it \textendash{} in the vv-list. The start value may be a \LaTeX{}
expression. Both \verb`rstep` and \verb`rstop` may also be \LaTeX{}
expressions. They are evaluated \emph{after} the vv-list is evaluated
and so may depend on the values of variables in the vv-list, including
the \emph{initial} value of the row variable. As of v.3.2 of \texttt{numerica-tables}
the \verb`vv@=1`\index{vv-list!vva@\texttt{vv"@} (mode) setting}
(or \verb`vvmode=1`) setting from \texttt{numerica} does not affect
\verb`rstep` and \verb`rstop`. For instance if \verb`rvar=x`, setting
\verb`rstep=1/x` when \verb`vv@=1` gives a \emph{constant} step
size equal to the reciprocal of the initial\emph{ }value of the row
variable. The following examples illustrate the use of \verb`rvar`,
\verb`rstep` and \verb`rstop` and the presence of the initial value
of the row variable in the vv-list.
\begin{verbatim}
    \tabulate[rvar=x,rstep=0.2,rstop=1]
      { \sin x }[x=0]\qquad
    \tabulate[rvar=x,rstep=0.2,rstop=1]
      { \sin x }[x=0][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rvar=x,rstep=0.2,rstop=1]
      { \sin x }[x=0] \qquad
    \tabulate[rvar=x,rstep=0.2,rstop=1]
      { \sin x }[x=0][*]\medskip{}

The difference in appearance of the tables results from padding with
zeros\index{function-value formatting!padding with zeros}\index{padding with zeros}
in the second. The asterisk in the trailing optional argument has
the same effect in \verb`\nmcTabulate` as in \verb`\nmcEvaluate`.
As you can see, padding applies not only to the values of the function
but also to the values of the row variable, although that has been
padded to only $1$ decimal place (the default) rather than the $6$
of the function values. (How many digits to round to, and therefore
pad the row variable values to, is discussed in §\ref{subsec:Rounding:rround}.)
Padding makes an obvious improvement to the appearance, both in the
function-value column and the row variable column. 

You may feel a single column of function values makes poor use of
space on the page. To turn the second of these tables into a `horizontal'
table, add the key \verb`transpose` to the settings option. I have
also reduced the number of decimal places from the default $6$ to
$4$ to avoid any issues with page width.
\begin{verbatim}
    \tabulate[rvar=x,rstep=0.2,rstop=1,transpose]
      { \sin x }[x=0][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rvar=x,rstep=0.2,rstop=1,transpose]
      { \sin x }[x=0][4*]

\medskip{}

\index{transpose@\texttt{transpose}!examples of use}To my eye, for
such a `shallow' (vertically compressed) table, the horizontal rules
here are intrusive. They can be turned off; see §\ref{subsec:Rules:-rules-setting}
or, anticipating the later discussion, add \verb`norules`\index{norules@\texttt{norules}}
to the settings option (prior to version 3.2 of \texttt{numerica-tables}
it was necessary to write \verb`rules=`\texttt{ }, an empty setting,
to achieve this), and just for variety use the cosine:
\begin{verbatim}
    \tabulate[rvar=x,rstep=0.2,rstop=1,transpose,norules]
      { \cos x }[x=0][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rvar=x,rstep=0.2,rstop=1,transpose,norules]
      { \cos x }[x=0][4*] \index{transpose@\texttt{transpose}!examples of use}

\medskip{}

The remaining quibble is the placement of the column headers. Can
they be centred or otherwise adjusted? Yes; see §\ref{subsec:Row-var-col-formatting}
for ways of adjusting the row variable column and header, and §\ref{subsec:Column-header-formatting}
for ways of adjusting the function-value columns and headers. 

It may be more convenient at times to specify the number of rows,
\verb`rows`\index{rows@\texttt{rows}}, to tabulate rather than a
stop value. Only one of \verb`rows` and \verb`rstop` should be given,
but if both (inadvertently) are present, it is  \verb`rows` that
prevails. The first of the following three tables shows an example
where \verb`rows` is specified. The second and third tables use an
abbreviated form of the row variable specification, \verb`rspec`\index{rspec@\texttt{rspec}}.
This is a three-element comma list, \verb`{rvar,rstep,rows}` (with
an optional fourth element, \verb`rround`; see §\ref{subsec:Rounding:rround}).
The second table gives a straightforward example. In the third table
a simple \LaTeX{} expression has been inserted for \verb`rows` in
the \verb`rspec` comma list. Like \verb`rstep` and \verb`rstop`,
\verb`rows` can be a \LaTeX{} expression but unlike \verb`rstep`
and \verb`rstop` it is evaluated \emph{before} the vv-list and therefore
cannot depend on quantites specified there like the initial row variable
value. On evaluation, the \LaTeX{} expression is rounded to an integer.
\begin{verbatim}
    \tabulate[rvar=x,rstep=0.2,rows=6]
      { \sin x/\cos x }[x=0][*] \qquad
    \tabulate[rspec={x,0.2,6}]
      { \tan x }[x=0][*] \qquad
    \tabulate[rspec={x,0.2,1.2/0.2}]
      { \sqrt{\sec^2 x - 1} }[x=0.2][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rvar=x,rstep=0.2,rows=6]
      { \sin x/\cos x }[x=0.2][*] \qquad
    \tabulate[rspec={x,0.2,6}]
      { \tan x }[x=0.2][*] \qquad
    \tabulate[rspec={x,0.2,1.2/0.2}]
      { \sqrt{\sec^2 x - 1} }[x=0.2][*]\index{row variable!start, step, stop values|)}

\subsubsection{Rounding: \texttt{rround}}

\label{subsec:Rounding:rround}After\index{rounding value!row variable|(}\index{rround@\texttt{rround}|(}
studying some of the previous tables, we might decide to adjust the
step size, say from $0.2$ to $0.25$. But changing \verb`rstep`
to the new value gives a disconcerting result, as you can see from
the first of the tables below. \texttt{numerica-tables} uses a default
rounding value of $1$ for the row variable and has rounded $0.25$
down to $0.2$, then $0.2+0.25=0.45$ down to $0.4$, then $0.4+0.25=0.65$
down to $0.6$, then $0.6+0.25=0.85$ down to $0.8$, at which point
it stops since $0.8+0.25>1$ which is the stopping value. The rounded-down
values are the values used \textendash{} always it is the \emph{displayed}
rounded value that is used for calculating function values. The second
table corrects matters by adjusting the row variable rounding with
\verb`rround=2`. From version 3.2 of \texttt{numerica-tables} the
\verb`rround` value can also be included as an optional fourth element
to \verb`rspec`\index{rspec@\texttt{rspec}}. The third table illustrates
this abbreviated form of specification:
\begin{verbatim}
    \tabulate[rvar=x,rstep=0.25,rstop=1]
      { \exp x }[x=0][*]  \qquad
    \tabulate[rvar=x,rstep=0.25,rstop=1,rround=2]
      { \exp x }[x=0][*] \qquad
    \tabulate[rspec={x,0.25,5,2}]
      { \exp x }[x=0][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rvar=x,rstep=0.25,rstop=1]
      { \exp x }[x=0][*] \qquad
    \tabulate[rvar=x,rstep=0.25,rstop=1,rround=2]
      { \exp x }[x=0][*] \qquad
    \tabulate[rspec={x,0.25,5,2}]
      { \exp x }[x=0][*]\medskip{}

The reason for allowing the inclusion of the \verb`rround` value
in \verb`rspec` is because the values included in \verb`rspec`,
together with the initial value assigned to the row variable in the
vv-list, determine the numbers in the row variable column used to
calculate function values. The further settings discussed below in
§\ref{subsec:Row-var-col-formatting} are \emph{formatting} elements
affecting appearance but irrelevant to the actual values calculated.\index{rround@\texttt{rround}|)}\index{rounding value!row variable|)}

\subsection{Row variable specification: non-uniform case}

\label{subsec:Row-var-spec-nonuniform}Occasionally one wants to form
a table in which the row variable does not increase or decrease in
regular steps; for examples, see \emph{HMF} Tables 1.1 and 3.1. (Tables
9.7, 10.5, 10.10 use two step values and could also be handled with
these settings.) For instance, one might want a table of values of
simple functions of a list of constants, or a table of function values
at $\pi$, $\pi/2$, $\pi/3$, $\pi/4$, $\dots$, or at $1$, $10$,
$100$, $1000$, $\dots$, or a table of function values at thoroughly
irregular, perhaps experimentally determined, values.

\texttt{numerica-tables} provides two means of specifying such row
variables, either by means of a row variable function (\verb`rfunc`\index{rfunc@\texttt{rfunc}}),
when the row variable values change in a non-uniform but formulaic
way, or by explicitly listing the row variable values in a comma list
(\verb`rdata`\index{rdata@\texttt{rdata}}, \verb`rfile`\index{rfile@\texttt{rfile}}),
or with some other separator. In the latter case, the row variable
can be displayed either as a sequence of \emph{values}, or verbatim
as a sequence of \emph{expressions} \textendash{} like fractions of
$\pi$ \textendash{} with the \verb`rverb`\index{rverb@\texttt{rverb}}
setting.

\subsubsection{\texttt{rfunc}}

\begin{table}[t]
\centering
\centering{}\caption{Non-uniform row variable specification}
\begin{center}
\begin{tabular}{l>{\raggedright}p{2cm}>{\raggedright}p{4cm}>{\raggedright}p{3cm}}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small comment}\tabularnewline
\midrule
{\small\texttt{rfunc}} & {\small token(s)} & {\small formula for row var. values} & \tabularnewline
{\small\texttt{rdata}} & {\small comma list or macro} & {\small list or macro (containing list) of row var. values} & \tabularnewline
{\small\texttt{rfile}} & {\small chars} & {\small filepath/filename} & {\small file contains list of row var. values}\tabularnewline
{\small\texttt{rdelim}} & {\small char} & {\small item separator for }{\small\texttt{rdata}}{\small{} or }{\small\texttt{rfile}}{\small{}
lists} & {\small defaults to }{\small\texttt{,}}{\small{} or }{\small\texttt{;}}{\small{}
depending as }{\small\texttt{.}}{\small{} or }{\small\texttt{,}}{\small{}
is decimal mark}\tabularnewline
{\small\texttt{rverb}} & {\small fp (}{\small\texttt{0/0.5/1}}{\small )} & {\small display }{\small\texttt{rdata}}{\small , }{\small\texttt{rfile}}{\small{}
values verbatim (}{\small\texttt{1}}{\small ), or slash fractions formatted
(}{\small\texttt{0.5}}{\small )} & {\small initialized to }{\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}
\index{rfunc@\texttt{rfunc}|(}\index{row variable!variably stepped|(}\label{subsec:rfunc}Suppose
\textendash{} perhaps with an interest in the distribution of prime
numbers \textendash{} that we want to create a small table of values
of $n/\ln n$ for, say, $n=10$, $100$, 1000, $10000$, $\ldots$
The prospective row variable $n$ is not increasing uniformly, although
clearly in a formulaic way. The key \verb`rfunc` is for such cases;
in the present instance \verb`rfunc=10^n` where now \verb`n` does
increment by a constant amount:
\begin{verbatim}
    \tabulate[rfunc=10^n,rvar=n,rstep=1,rows=7] 
      { n/\ln n }[n=1][0] \qquad
    \tabulate[rfunc=10^n,rvar=n,rstep=1,rows=7,norules] 
      { n/\ln n }[n=1][0]
\end{verbatim}
$\Longrightarrow$     \tabulate[rpos=1,rfunc=10^n,rvar=n,rstep=1,rows=7]
     { n/\ln n }[n=1][0]\qquad
    \tabulate[rfunc=10^n,rvar=n,rstep=1,rows=7,norules] 
      { n/\ln n }[n=1][0]\medskip{}

The benefit of the \verb`norules`\index{norules@\texttt{norules}}
setting in the second table is obvious.

The variable \verb`n` has two meanings in these tables although only
one is publicly visible. Initially \verb`n` is the variable of a
step function, incremented by \verb`rstep=1` and taking values \verb`10^n`.
Once the table is compiled, \verb`n` denotes these successive function
values, $10,100,\ldots,10000000$. To the \emph{reader}, only this
latter meaning is evident, the potentially confusing `double usage'
not so. The initial value \verb`n=1`\index{vv-list!row/col.variables@row/col.\ variables}\index{row variable!initial value in vv-list}
in the vv-list applies to the row variable function \verb`10^n`,
not to the function \verb`n/\ln n` being tabulated (so the error-producing
expression \verb`1/\ln 1` does not arise).\index{rfunc@\texttt{rfunc}|)}\index{row variable!variably stepped|)}

\subsubsection{\texttt{rdata}, \texttt{rfile}, \texttt{rdelim}, \texttt{rverb} }

\label{subsec:rdata-rfile-rverb}\index{row variable!values from list/macro/file|(}A
difficulty in reading the last two tables is working out just how
many zeros there are in the larger numbers in the left column. They
would be more readable `at a glance' if we could write them in scientific
notation. To do that, use the \verb`rdata`\index{rdata@\texttt{rdata}|(}
and \verb`rverb`\index{rverb@\texttt{rverb}|(} keys. In the tables
above, \verb`rverb` is absent (corresponding to \verb`rverb=0`);
in the tables below \verb`rverb=1`, the effect of which is to use
the row variable values in the \verb`rdata` list verbatim.\index{row variable!verbatim values|(}
In the first example below, the larger values of the row variable
are now expressed in scientific notation, which gives a table that
is easier to grasp \textendash{} but less pleasing to the eye. The
right alignment\index{ralign@\texttt{ralign}} of the column is the
culprit, as is seen when presented left-aligned in §\ref{subsec:RValignment:ralign}.
\begin{verbatim}
    \tabulate[rdata={10,100,1000,10^4,10^5,10^6,10^7},
        rverb=1,rvar=n]{ n/\ln n }[0] \qquad 
    \def\mydata{\sfrac14\,\pi,\sfrac13\,\pi,\sfrac12\,\pi,
                 \sfrac23\,\pi,\sfrac34\,\pi,\pi}
    \tabulate[rdata=\mydata,rverb=1,rvar=k] 
      { k }[*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rdata={10,100,1000,10^4,10^5,10^6,10^7},
        rverb=1,rvar=n,rround=0]{ n/\ln n }[0] \qquad 
    \def\mydata{\sfrac14\,\pi,\sfrac13\,\pi,\sfrac12\,\pi,
                 \sfrac23\,\pi,\sfrac34\,\pi,\pi}
    \tabulate[rdata=\mydata,rverb=1,rvar=k] 
      { k }[*]  \medskip{}

In the second example \verb`rverb=1` is used to make a table of simple
fractions of $\pi$, listed verbatim in the row variable column against
their decimal values in the second column (the pointless header row
is discussed shortly). When \verb`rverb=1`, \verb`rround`\index{rround@\texttt{rround}}
does not apply. The full value is used in the calculation. For the
fractions I have used the \verb`\sfrac`\index{sfrac@\texttt{\textbackslash sfrac}}
command from the \texttt{xfrac}\index{xfrac package@\texttt{xfrac} package}
package (loaded in the preamble to the present document). \verb`\sfrac`
produces an elegant inline fraction with a smaller vertical footprint
than \verb`\tfrac` and so is better suited to use in a table column
that aligns fraction above fraction. The data has been stored in the
macro \verb`\mydata`. By setting \verb`rdata` equal to this macro,
the \verb`\tabulate` command gains access to the values stored in
it. 

In addition to a list or a macro, data for the row variable can also
be stored as a list in a file \textendash{} say \verb`mydata.txt`.
If \verb`mydata.txt` is placed in the directory of the current document
and \verb`rfile=mydata.txt`\index{rfile@\texttt{rfile}} entered
in the settings option of the \verb`\tabulate` command, the file
will be found and the values in the file used for the row variable.
Alternatively, the file could be placed in your \verb`texmf` tree
and your \TeX{} distro alerted to its presence (by refreshing the filename
database). Again \verb`rfile=mydata.txt` in the settings option will
ensure the file is found and the contents used for the row variable
values. Or, the file could be stored elsewhere and the \verb`rfile`
key equated to the full path and filename \textendash{} something
like \verb`rfile=e:/mydocs/mydatafiles/mydata.txt`. This ensures
the file will be found and the contents used for the row variable
values. Note that even in Windows systems (where file paths use backslashes)
the path requires that forward slashes only be used. \index{row variable!values from list/macro/file|)}

The example above uses commas to separate the values accessed by \verb`rdata`.
In earlier versions of \texttt{numerica-tables} this was the case
even if the \verb`comma` package option was used with \texttt{numerica}
(meaning values using a comma as the decimal mark would need to be
enclosed in braces). From version 3.2 of \texttt{numerica-tables},
if the \verb`comma` package option\index{comma package option@\texttt{comma} package option}
is used when calling \texttt{numerica}, then the list that \verb`rdata`
is equated to (or the file that \verb`rfile` references) should be,
by default, a \emph{semicolon} list.\index{decimal comma!semicolon use}\index{decimal comma!and rdata, rfile@and \texttt{rdata}, \texttt{rfile}}
It is also possible, from version 3.2, to specify your own delimiter
(item separator) by using the \verb`rdelim`\index{rdelim@\texttt{rdelim}}
setting. For instance, \verb`rdelim=|` would mean that the `pipe'
character separates items in the \verb`rdata` list (or the file referenced
by \verb`rfile`).

An irritant with the second table above is the pointless header row.
To suppress it I could anticipate and enter \verb`headless`\index{headless@\texttt{headless}}
(see §\ref{subsec:Header-row-suppress}) in the settings option. But
perhaps a better solution is to present the table horizontally by
means of the \verb`transpose` setting. Now the pointless header in
the vertical table becomes a pointless first column in the horizontal
table. The way to suppress that is by giving a zero value to the setting
\verb`rpos`\index{rpos@\texttt{rpos}} (see §\ref{subsec:Row-var-col-pos}
for a discussion of \verb`rpos`). \verb`rpos` determines the placement
of the row variable column and a zero value suppresses its display
entirely. I have also suppressed all rules with the \verb`norules`\index{norules@\texttt{norules}}
setting, and reduced the rounding value to $4$. At which point the
remaining quibble is with the default right alignment of the entries
in the header row. I have anticipated again and centred each column
with the \verb`calign=c`\index{calign@\texttt{calign}} setting (§\ref{subsec:Alignment:-calign}).
\begin{verbatim}
    \def\mydata{\sfrac14\,\pi, \sfrac13\,\pi, \sfrac12\,\pi,
                \sfrac23\,\pi, \sfrac34\,\pi, \pi}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,rpos=0,calign=c,
                norules,transpose]{ k }[4*]
\end{verbatim}
$\Longrightarrow$     \def\mydata{\sfrac14\,\pi, \sfrac13\,\pi, \sfrac12\,\pi,
                \sfrac23\,\pi, \sfrac34\,\pi, \pi }
    \tabulate[rdata=\mydata,rverb=1,rvar=k,rpos=0,calign=c,
                norules,transpose]{ k }[4*] \index{transpose@\texttt{transpose}!examples of use}\index{norules@\texttt{norules}}\medskip{}


\subsubsection{Fraction-form values}

\label{subsec:rfracFraction-form-rvar}\index{row variable!fraction form|(}\index{fraction form!row variable values|(}
The \verb`rverb` setting applies \emph{only} to the \verb`rdata`
and \verb`rfile` keys. It has no effect otherwise. Besides \verb`0`
and \verb`1`, it can also take a third value, \verb`rverb=1/2`,
which can be thought of as `semi-verbatim'\index{rverb@\texttt{rverb}!`semi-verbatim'}.
(\verb`rverb=0.5` or indeed any expression in the syntax of \texttt{l3fp}
\textendash{} like \verb`sin(pi/6)` \textendash{} that evaluates
to one half will serve as well.) This renders `naked' slash fractions
from \verb`rdata` or \verb`rfile` input in formatted form using
the \verb`\sfrac` command from the \texttt{xfrac} package if it is
loaded (e.g. $\sfrac 23$), or in scriptstyle (e.g. $\scriptstyle 2/3$)
if it isn't. A `naked' slash fraction here means the fraction isn't
buried within parentheses or other constructs.

I repeat the vertical presentation of the last table but use the `semi-verbatim'
setting. By moving the repeated factor \verb`\pi` from the data file
to the function being tabulated, the data file becomes a list of `naked'
slash fractions (and one integer) and the header row becomes meaningful.
Note that the integer is displayed as is with the `semi-verbatim'
setting; only the slash fractions are reformatted. Again, I have centred
the second column with the \verb`calign=c` setting:
\begin{verbatim}
    \def\mydata{1/4,1/3,1/2,2/3,3/4,1}
    \tabulate[rdata=\mydata,rverb=1/2,
              rvar=k,calign=c] 
      { k\pi }[k=0][6*]
\end{verbatim}
$\Longrightarrow$     \def\mydata{1/4,1/3,1/2,2/3,3/4,1}
    \tabulate[rdata=\mydata,rverb=0.5, 
        rvar=k,rhnudge=6,chnudge=18] 
     { k\pi }[k=0][6*]

\subsection{Formatting the row variable column \& header}

\label{subsec:Row-var-col-formatting}The padding option ({*}) of
the trailing optional argument is one way of formatting the row variable
values, but to how many decimal places? Aligned left or right or centred?
Under what heading \textendash{} the examples so far have simply used
the row variable for the header? And should the row variable column
be at the left of the table, or the right \textendash{} or both? These
and related questions are answered by assigning values to the keys
listed in Table~\ref{tab:Formatting-the-row-variable}.

\begin{table}[t]
\centering
\centering{}\caption{\label{tab:Formatting-the-row-variable}Formatting the row variable
column \& header}
\begin{center}
\begin{tabular}{l>{\raggedright}p{2cm}>{\raggedright}p{4cm}>{\raggedright}p{2cm}}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small initial}\tabularnewline
\midrule
{\small\texttt{rpos}} & {\small int (}{\small\texttt{0}}{\small\ldots}{\small\texttt{4}}{\small )} & {\small column placement} & {\small\texttt{1}}\tabularnewline
{\small\texttt{ralign}} & {\small char (}{\small\texttt{r/c/l}}{\small )} & {\small horizontal alignment} & {\small\texttt{r}}\tabularnewline
{\small\texttt{rfont}} & {\small chars} & {\small font (}{\small\verb`\math<chars>`}{\small )} & \tabularnewline
{\small\texttt{rhead}} & {\small tokens} & {\small header} & \tabularnewline
{\small\texttt{rhnudge}} & {\small fp} & {\small nudge header }{\small{\small\verb`<fp>`}}{\small{} mu} & {\small\texttt{0}}\tabularnewline
{\small\texttt{rhsize}} & {\small int (}{\small\texttt{-4}}{\small\ldots}{\small\texttt{5}}{\small )} & {\small font size relative to }{\small\verb`0=\normalsize`} & {\small\texttt{0}}\tabularnewline
{\small\texttt{rmath}} & {\small char (}{\small\texttt{s/t/d}}{\small )} & {\small script-, text-, displaystyle} & {\small\texttt{t}}\tabularnewline
{\small\texttt{rvar'}} & {\small tokens} & {\small 2nd row variable col. spec. } & \tabularnewline
{\small\texttt{rhead'}} & {\small tokens} & {\small header of 2nd row var. col. (if it exists) } & \tabularnewline
{\small\texttt{rhnudge'}} & {\small fp} & {\small nudge 2nd row var. col. header }{\small{\small\verb`<fp>`}}{\small{}
mu} & {\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}


\subsubsection{Position in the table: \texttt{rpos}}

\label{subsec:Row-var-col-pos}\index{row variable column!position in table}\index{rpos@\texttt{rpos}}By
default, the row variable column is the \emph{first} column of the
table. Its position is determined by the value of the key \texttt{rpos}: 
\begin{itemize}
\item \verb`rpos=0`, suppressed (no row variable column); 
\item \verb`rpos=1`, first column (the default); 
\item \verb`rpos=2`, last column; 
\item \verb`rpos=3`, first and last columns (e.g. see §\ref{sec:Multiple-function-tables});\texttt{ }
\item \verb`rpos=4`, first column and a last column with values produced
by a user-defined function of the first; see §\ref{subsec:Second-row-var-col};
\item Any other integer acts like \verb`rpos=1`.
\end{itemize}
When a table is transposed, these settings apply to the \emph{new}
row variable column (the old header row).\index{transpose@\texttt{transpose}!and rpos@and \texttt{rpos}}
For instance, with \verb`rpos=2` we get what was originally the header
row treated as the row variable column and placed on the right:
\begin{verbatim}
     \tabulate[rspec={x,0.2,5},rpos=2,transpose,norules]
         { \tan x }[x=0.2][*]
\end{verbatim}
$\Longrightarrow$      \tabulate[rspec={x,0.2,5},rpos=2,transpose,norules]
         { \tan x }[x=0.2][*]\index{transpose@\texttt{transpose}!examples of use}\index{norules@\texttt{norules}}

\paragraph{Adjoined multi-function tables}

\label{subsec:Adjoining-tables}\index{adjoined tables}\index{multi-function tables}
By assigning different values to \verb`rpos` it is possible to include
multiple functions in the same table by adjoining distinct single
column tables. \emph{HMF} has many, many examples where multiple functions
(like the trigonometric or hyperbolic functions) are tabulated in
separate columns of the same table. 

\texttt{numerica-tables} has a systematic way of doing this, described
in §\ref{sec:Multiple-function-tables}, but with the settings discussed
to this point, the same effect can be achieved by the present more
naive means. The example below displays as a single multi-columned
table but is composed of three separate tables. I have used three
different\index{rpos@\texttt{rpos}} \verb`rpos` settings (\verb`rpos=1`,
the default setting, is implicit in the first) to achieve this and
ended all but the last \verb`\tabulate` command with the \LaTeX{}
comment character \verb`%`. These characters are essential if the
tables are to abut exactly. Omitting them results in a space between
the tables.
\begin{verbatim}
    \tabulate[rspec={x,0.2,6}]
      { \sin x }[x=0][*]%
    \tabulate[rpos=0,rspec={x,0.2,6}]
      { \cos x }[x=0][*]%
    \tabulate[rpos=2,rspec={x,0.2,6}]
      { \tan x }[x=0][*]
\end{verbatim}
$\Longrightarrow$     \tabulate
      [rspec={x,0.2,6}]
      { \sin x }[x=0][*]%
    \tabulate
      [rpos=0,rspec={x,0.2,6}]
      { \cos x }[x=0][*]%
    \tabulate
      [rpos=2,rspec={x,0.2,6}]
      { \tan x }[x=0][*]\medskip{}

It is possible to similarly join transposed tables (you need the \verb`headless`
setting), stacked one below the other, but again that effect can be
achieved more readily in a systematic way \textendash{} see §\ref{sec:Multiple-function-tables}.

\subsubsection{Alignment: \texttt{ralign}}

\label{subsec:RValignment:ralign}\index{row variable column!alignment|(}\index{ralign@\texttt{ralign}}By
default, the alignment of all columns including the row variable column
is to the right, as in the examples so far. This lends itself to clean
output when padding with zeros is activated (the \verb`*` in the
trailing argument) or when some values are negative, since minus signs
can interfere with alignment of digits in left or centred alignments.
In the first example below I present an earlier table where the row
variable values were a mix of powers of $10$ multiplied out and in
exponent form. In that table (see §\ref{subsec:rdata-rfile-rverb})
with its right alignment the eye stumbled over the exponents. In left
alignment, it is a column of `$1$'s which is aligned and the alignment
is `smooth'. Left alignment is achieved by setting \verb`ralign=l`
(a lowercase `L'). In the second table I have chosen a centred alignment
for the row variable column, \verb`ralign=c` (and once more used
\verb`calign=c` for the function-value column). The other possible
setting for \verb`ralign` is the default right alignment, \verb`ralign=r`.
\begin{verbatim}
    \tabulate[rdata={10,100,1000,10^4,10^5,10^6,10^7},
        rverb=1,rvar=n,ralign=l] 
      { n/\ln n }[0]\qquad
     \tabulate[rspec={x,0.01,7},rround=2,ralign=c,calign=c]
       { \sin x }[x=1.51][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rdata={10,100,1000,10^4,10^5,10^6,10^7},
        rverb=1,rvar=n,ralign=l] 
     { n/\ln n }[0]\qquad
     \tabulate[rspec={x,0.01,7},rround=2,
                 ralign=c,calign=c]
       { \sin x }[x=1.51][*]

\subsubsection{Font: \texttt{rfont}}

\index{rfont@\texttt{rfont}}\index{formatting!row variable column|(}You
may wish to distinguish the values in the row variable column from
the values in the function-value column typographically, for instance
by bolding them. This is effected through the setting \verb`rfont`.
Possible values for this setting are \verb`rm` (roman), \verb`bf`
(bold face), \verb`it` (italic), \verb`sf` (sans serif), and \verb`tt`
(typewriter). For instance, \verb`rfont=bf` applies \verb`\mathbf`
to each row variable value, \verb`rfont=tt` applies \verb`\mathtt`
to each row variable value, and so on. If you set \verb`rfont=xx`
when \verb`\mathxx` has not been defined, then a \LaTeX{} error will
result. Equally, something like \verb`rfont=cal` (where \verb`\mathcal`
\emph{has} been defined) should not cause error but will give unexpected
results if the digits $0$ to $9$ are not part of the font. In the
next example, I repeat the last example in `vertical' form but with
the row variable column values bolded to draw attention to their placement
on the right \textendash{} but note that the \verb`rfont` setting
has no effect on the row variable header.
\begin{verbatim}
     \tabulate[rspec={x,0.2,5},rpos=2,rfont=bf]
         { \tan x }[x=0.2][*] \qquad
     \tabulate[rspec={x,0.2,5},rpos=2,rfont=bf,
                 rhead=\boldsymbol{x}\phantom{.}]
         { \tan x }[x=-0.4][*]
\end{verbatim}
$\Longrightarrow$      \tabulate[rspec={x,0.2,5},rpos=2,rfont=bf]
         { \tan x }[x=0.2][*] \qquad
     \tabulate[rspec={x,0.2,5},rpos=2,rfont=bf, 
                 rhead=\boldsymbol{x}\phantom{.}]
         { \tan x }[x=-0.4][*]

\subsubsection{Row variable header: \texttt{rhead}}

\label{subsec:RVheader:rhead}\index{rhead@\texttt{rhead}|(}\index{row variable header|(}The
default header is the row variable symbol. In the second table that
has been changed  by giving a value to the key \verb`rhead`. I have
used \verb`rhead=\boldsymbol{x}` (rather than \verb`\mathbf{x}`)
in order to get an italicized bold symbol. I've also added a \verb`\phantom{.}`
to \verb`rhead` in order to push the (by default) right-aligned header
leftwards and give a more centred appearance to its positioning. Math
delimiters (\verb`$ $`) are inserted automatically by \texttt{numerica-tables}.
If you do not want any header at all for the row variable column,
entering \verb`rhead= ` , an empty setting, will achieve this \textendash{}
at least from version 3.2 of \texttt{numerica-tables}. In earlier
versions you needed to set \verb`rhead` to (e.g.) a space of some
kind like \verb`\ ` (an inter-word space). \index{rhead@\texttt{rhead}|)}\index{formatting!row variable header|)}

\subsubsection{Nudging the header: \texttt{rhnudge}}

\label{subsec:RVnudging:rhnudge}Choosing a centred alignment for
the header to the row variable column will generally not be appropriate
when negative values are involved. In an earlier table I used a phantom
to get around this. Another way which avoids obscuring the true content
of the \verb`rhead` setting with positioning commands is to use the
setting \verb`rhnudge`\index{rhnudge@\texttt{rhnudge}|(}. 

Both tables below have the default right alignment of the row variable
column and use \verb`rhnudge` to adjust the header position. For
positive values, \verb`rhnudge` works in the \emph{opposite} sense
to the alignment, to the left for right alignment, and to the right
otherwise. The units for nudging are mu (math units, 18 to a quad),
but only a number  should be specified; the `mu' is supplied by
\texttt{numerica-tables}.
\begin{verbatim}
    \tabulate[rvar=x,rstep=0.25,rstop=0.5,rround=2,
              rfont=bf,rhead=\boldsymbol{x},rhnudge=9]
      { \sin x }[x=-0.5][4*]\qquad
    \tabulate[rvar=x_{\text{int}},rstep=1,rstop=4,
              rround=0,rfont=bf,rhnudge=-12,
              rhead=\boldsymbol{x_{\text{int}}}]
      { \exp x_{\text{int}} }[x_{\text{int}}=0][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate
      [rvar=x,rstep=0.25,rstop=0.5,rround=2,
        rfont=bf,rhead=\boldsymbol{x},rhnudge=9]
      { \sin x }[x=-0.5][4*]\qquad
    \tabulate
      [rvar=x_{\text{int}},rstep=1,rstop=4,
        rround=0,rfont=bf,rhnudge=-12,
        rhead=\boldsymbol{x_{\text{int}}}]
      { \exp x_{\text{int}} }[x_{\text{int}}=0][4*]\medskip{}

\noindent In the second table the row variable takes single digit
integer values, while the row variable name occupies more than one
character. With a right alignment the header would protrude out to
the left. By giving \verb`rhnudge` a \emph{negative} value (\verb`rhnudge=-12`
in the example) it is brought back to a centred position in the row
variable column. \index{rhnudge@\texttt{rhnudge}|)}\index{formatting!row variable column|)}

\subsubsection{Font size of the header: \texttt{rhsize}}

The font size of the header can be changed by assigning an integer
value to the key \verb`rhsize`\index{rhsize@\texttt{rhsize}}. Valid
values range from $-4$, corresponding to \verb`\tiny`, through $0$,
corresponding to \verb`\normalsize`, to $5$, corresponding to \verb`\Huge`.
The actual values in points differ depending as \verb`\normalsize`
is $10$pt, $11$pt, or $12$pt, but range from $5$pt for \verb`\tiny`
when \verb`\normalsize` is $10$pt, to $24.88$pt for \verb`\Huge`
in all three cases. See also §\ref{subsec:chsize}.

\subsubsection{\texttt{rmath}}

\label{subsec:rmath}By\index{rmath@\texttt{rmath}} default the row
variable column is displayed in textstyle (\verb`\textstyle`). By
setting \verb`rmath` to \verb`s`, \verb`t` or \verb`d` it is possible
to explicitly choose \verb`\scriptstyle`, \verb`\textstyle` or \verb`\displaystyle`
for its display. \verb`rmath` is relevant when one might want to
display more complicated structures than numbers in the row variable
column, e.g. verbatim items when using \verb`rdata` or \verb`rfile`,
or in transposed multi-function tables; see the example in §\ref{sec:Multiple-function-tables}.
Returning to an earlier example, suppose the macro \verb`\mydata`
contains a list of fractions, not in slash form but \verb`\frac`-form:\index{arraystretch@\texttt{\textbackslash arraystretch}}
\begin{verbatim}
    \def\mydata{\frac14,\frac13,\frac12,\frac23,\frac34,1}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,calign=c]
      { k\pi }[k=0][6*]\qquad
    \renewcommand\arraystretch{1.2}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,calign=c]
      { k\pi }[k=0][6*]\qquad
    \renewcommand\arraystretch{1}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,rmath=s,calign=c]
      { k\pi }[k=0][6*]
\end{verbatim}
$\Longrightarrow$     \def\mydata{\frac14,\frac13,\frac12,\frac23,
                \frac34,1}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,calign=c]
      { k\pi }[k=0][6*]\qquad
    \renewcommand\arraystretch{1.2}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,calign=c]
      { k\pi }[k=0][6*]\qquad
    \renewcommand\arraystretch{1}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,rmath=s,calign=c]
                     { k\pi }[k=0][6*]\medskip{}

\noindent The first table is unacceptable; the second (with \verb`\arraystretch`
set to $1.2$) and the third (with \verb`rmath=s`) are at least readable
but neither is really satisfactory. Better is to use \verb`\sfrac`,
either directly or by the `semi-verbatim' \verb`rverb` setting
discussed earlier in §\ref{subsec:rfracFraction-form-rvar}.\index{fraction form!row variable values|)}\index{row variable!fraction form|)}\index{row variable!verbatim values|)}\index{rverb@\texttt{rverb}|)}\index{rdata@\texttt{rdata}|)}

\subsubsection{\texttt{rvar'}, \texttt{rhead'}, \texttt{rhnudge'}}

These settings become relevant only when \verb`rpos=4`; see §\ref{subsec:Second-row-var-col}.\index{row variable header|)}

\section{Column-variable settings}

\label{sec:Column-variable-settings}When a function of \emph{two}
variables is tabulated, we generally think of one variable as the
primary variable and the other as a parameter. To tabulate such a
function, we use the primary variable as the row variable and create
a succession of adjacent columns of function values according to the
different values of the parameter. In this document I call this second
variable the \emph{column} variable (\verb`cvar`). Just as with the
row variable, to create a table we need its start value (specified
in the vv-list), and its step (\verb`cstep`) and stop (\verb`cstop`)
values\index{column variable!start, step, stop values|(}. 
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cstop=9]
      { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cstop=9]
      { \sin kx}[k=3,x=0.2][4*] \medskip{}

In \index{cvar@\texttt{cvar}|(} this example \verb`cvar=k` is the
column variable. I have chosen a step size \verb`cstep=2`\index{cstep@\texttt{cstep}|(}
and a stop value \verb`cstop=9`\index{cstop@\texttt{cstop}|(}. The
start value ({\ttfamily\verb`k=3`}) is specified in the
vv-list\index{vv-list!row/col.variables@row/col.\ variables}\index{column variable!initial value in vv-list}.
Although in the example these values are numbers, all three values
could be \LaTeX{} expressions that evaluate to numbers. In particular,
the expressions for step and stop values may include the row and column
variables (in the example $x$ and $k$) which are assigned their
\emph{initial} vv-list values. 

\begin{table}[t]
\centering
\caption{Column-variable specification}

\centering{}\begin{center}
\begin{tabular}{ll>{\raggedright}p{3cm}l}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small comment}\tabularnewline
\midrule
{\small\texttt{cvar}} & {\small token(s)} & {\small column variable} & \tabularnewline
{\small\texttt{cstep}} & {\small real num.} & {\small step size} & \tabularnewline
{\small\texttt{cstop}} & {\small real num.} & {\small stop value} & {\small either }{\small\texttt{cstop}}\tabularnewline
{\small\texttt{cols}} & {\small int} & {\small number of columns} & {\small or }{\small\texttt{cols}}\tabularnewline
{\small\texttt{cspec}} & {\small comma list} & {\small\texttt{\{cvar,cstep}}{\small ,}{\small\texttt{cols\}}} & {\small short form spec.}\tabularnewline
{\small\texttt{chround}} & {\small int} & {\small col. var. rounding} & {\small default: }{\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}
\texttt{numerica-tables} has automatically inserted the row variable
header which now has a double function: it serves also as the `header'
of the header row. As you can see $x\mkern1mu\backslash\mkern1mu k$
has been inserted into this cell. From version 3.2 of \texttt{numerica-tables}
the automatic insertion of this form of header in a multi-column table
occurs if \verb`rhead`\index{rhead@\texttt{rhead}} is omitted. (If
you want nothing to show, put \verb`rhead= `, an empty setting.)
This form of header above the row variable column occurs throughout
\emph{HMF}. In the present example it shows the reader that the numerical
values displayed in the row variable column are values of $x$ and
the values in the column headers are values of $k$. 

As with the row variable, rather than using a stop value,\texttt{
}{\ttfamily\verb`cstop`}, you can specify the number of
columns, \verb`cols`\index{cols@\texttt{cols}|(}, explicitly. I
could have replaced {\ttfamily\verb`cstop=9`} with \verb`cols=4`
to get the same result. \verb`cols` specifies the number of \emph{function-value}
columns; the row variable column is ignored for this count. A quirk
of using \verb`cols` is that it is then possible to have a \emph{zero}
step size, \verb`cstep=0`.\index{cstep@\texttt{cstep}!zero value}
(A similar comment applies to \verb`rows` and \verb`rstep`\index{rstep@\texttt{rstep}!zero value}.)
For an instance where this is useful, see the last example in §\ref{subsec:Nudgingtheheaders:-chnudge}.

Again as with the row variable, the column specification can be condensed
into a comma list with the key \verb`cspec`\index{cspec@\texttt{cspec}|(}.
This is a $3$-element comma list of the form \verb`{cvar,cstep,cols}`
(and as for \verb`rspec`, with an optional fourth element, \verb`chround`;
see §\ref{subsec:chround}). Thus, for the preceding table I could
also have written
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cols=4]
      { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
or more succinctly
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},cspec={k,2,4}]
      { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
and produced the same table. 

Like their row equivalents, \verb`cstep`, \verb`cstop` and \verb`cols`
can all be \LaTeX{} expressions, and like those equivalents, the first
two are evaluated \emph{after }the vv-list and so may depend not only
on numbers and constants but also on the \emph{initial} values of
the row and column variables, which are assigned in the vv-list.\index{vv-list!row/col.variables@row/col.\ variables}
\verb`cols` is evaluated \emph{before }the vv-list; it may be a \LaTeX{}
expression but cannot depend on the row or column variable.\index{cvar@\texttt{cvar}|)}\index{cstep@\texttt{cstep}|)}\index{cstop@\texttt{cstop}|)}\index{cols@\texttt{cols}|)}\index{cspec@\texttt{cspec}|)}\index{column variable!start, step, stop values|)}

Multi-column tables can be transposed: simply add \verb`transpose`
to the settings option. I have done this for the last example. Transposition
automatically changes the positions of $k$ and $x$ in the new row
variable header:\index{transpose@\texttt{transpose}!examples of use}
\begin{verbatim}
     \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cstop=9,
             transpose]{ \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cstop=9,
                transpose]
      { \sin kx}[k=3,x=0.2][4*] \medskip{}

\noindent To my eye, this table looks odd. Admittedly, the oddness
is exaggerated by the new row variable column being a column of integers;
\verb`ralign=c` would help there. But the underlying problem remains:
we expect the coarser graining in the column variable and its header
row, and the finer graining in the row variable and its column. Transposition
has confounded this expectation.

\subsection{Rounding: \texttt{chround}}

\label{subsec:chround}\index{chround@\texttt{chround}|(}\index{rounding value!header row|(}Generally
the column variable increments in larger steps than the row variable.
Often these are integer steps. For that reason the default rounding
value of the column variable is $0$ (as against $1$ for the row
variable), but it can be changed by assigning a different value to
the key \verb`chround`. 

For transposed tables, rounding is applied \emph{before }transposition.
The function values displayed in the body of the table are calculated
for the \emph{displayed} values of the row and column variables. Transposition
doesn't change the displayed function values and therefore shouldn't
change the displayed row and column variable values, irrespective
of their placement in the table. In the last (transposed) example
the `new' column headers show that the original row variable rounding
has been retained. As an essential part of the column variable specification,
\verb`chround` can be included as an optional fourth element in \verb`cspec`\index{cspec@\texttt{cspec}}. 

Since the default column variable rounding is $0$, any non-integer
increment to that variable needs an explicit value to be assigned
to \verb`chround`. For instance if $k$ increments by, say, $0.25$,
then \verb`chround=2` will need to be entered in the settings option
of the \verb`\tabulate` command, or \verb`2` entered as the fourth
element of \verb`cspec`:
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},ralign=c,
              cspec={k,0.25,4,2}]
      { \sin kx }[k=3,x=0.2][*4]
\end{verbatim}
$\Longrightarrow$     \tabulate 
      [rspec={x,0.2,5},ralign=c,
         cspec={k,0.25,4,2}]
      { \sin kx }[k=3,x=0.2][*4]\index{rounding value!header row|)}\index{chround@\texttt{chround}|)}\index{formatting!column variable header}

\subsection{Column header formatting}

\label{subsec:Column-header-formatting}\index{formatting!column variable header|(}\index{column variable header!formatting|(}
\begin{table}[t]
\centering
\centering{}\caption{Formatting the column variable header}
\begin{center}
{\small{}%
\begin{tabular}{ll>{\raggedright}p{4cm}l}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small default}\tabularnewline
\midrule
{\small\texttt{chstyle}} & {\small int (}{\small\texttt{0}}{\small\ldots}{\small\texttt{4}}{\small )} & {\small header style} & {\small\texttt{0}}\tabularnewline
{\small\texttt{chead}} & {\small token(s)} & {\small user-defined col. var. header} & \tabularnewline
{\small\texttt{calign}} & {\small char (}{\small\texttt{r}}{\small /}{\small\texttt{c}}{\small /}{\small\texttt{l}}{\small )} & {\small column alignment} & {\small\texttt{r}}\tabularnewline
{\small\texttt{chnudge}} & {\small fp} & {\small nudge header }{\small{\small\verb`chnudge`}}{\small{}
mu} & {\small\texttt{0}}\tabularnewline
{\small\texttt{chfont}} & {\small chars} & {\small font (}{\small\verb`\math<chars>`}{\small )} & \tabularnewline
{\small\texttt{chmath}} & {\small char (}{\small\texttt{s}}{\small /}{\small\texttt{t}}{\small /}{\small\texttt{d}}{\small )} & {\small script-, text- or displaystyle} & {\small\texttt{t}}\tabularnewline
{\small\texttt{chsize}} & {\small int (}{\small\texttt{-4}}{\small\ldots}{\small\texttt{5}}{\small )} & {\small font size relative to }{\small\verb`0 => \normalsize`} & {\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}}
\par\end{center}
\end{table}
Column header formatting depends on whether the table has a single
column of function values, or multiple columns of function values.
For the latter there are a number of built-in style settings for the
header, accessed by assigning a value (\verb`0`, \verb`1`, \verb`2`,
\verb`3`) to the setting \verb`chstyle`. This can (sometimes) be
adjusted with the settings \verb`chfont`, \verb`chnudge` and \verb`chmath`.
If these built-in styles don't satisfy then it is possible to define
your own header to the function-value columns by using the setting
\verb`chead`. 

\subsubsection{Single-column header}

\index{column variable header!single column case}When there is only
one column of function values, the function being tabulated is by
default set as the header to the column. This can be nudged left or
right by giving a numerical value to \verb`chnudge`. It can be set
in script-, text- or displaystyle (see the example) by setting \verb`chmath`\index{chmath@\texttt{chmath}}
to \verb`s`, \verb`t` (the default) or \verb`d`. \verb`chfont`,
however, has no effect in the single column case. 
\begin{verbatim}
    \tabulate[rspec={x,0.25,5},rround=2,chmath=s]
       { \frac{2}{\sqrt\pi} e^{-x^2} }[x=0][*] \qquad
    \tabulate[rspec={x,0.25,5},rround=2]
       { \frac{2}{\sqrt\pi} e^{-x^2} }[x=1][*] \qquad
    \tabulate[rspec={x,0.25,5},rround=2,chmath=d]
       { \frac{2}{\sqrt\pi} e^{-x^2} }[x=2][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.25,5},rround=2,chmath=s]
       { \frac{2}{\sqrt\pi} e^{-x^2} }[x=0][*]\qquad
     \tabulate[rspec={x,0.25,5},rround=2]
       { \frac{2}{\sqrt\pi} e^{-x^2} }[x=1][*]\qquad
     \tabulate[rspec={x,0.25,5},rround=2,chmath=d]
       { \frac{2}{\sqrt\pi} e^{-x^2} }[x=2][*]\medskip{}

You may want some other header to the function-value column, one of
your own choosing. In that case give \verb`chead`\index{chead@\texttt{chead}}
a value. You are responsible for the entire content, including any
math environment; see the examples at §\ref{subsec:Fraction-form-output}.

\subsubsection{Multi-column header}

\label{subsec:Headerstyle:-multi-column-case}\index{column variable header!multi-column|(}The
default header style lists the column variable \emph{value} at the
head of each function-value column. Implicitly it corresponds to \verb`chstyle=0`.
\begin{verbatim}
     \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cstop=9]
       { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},cvar=k,cstep=2,cstop=9]
      { \sin kx}[k=3,x=0.2][4*]\medskip{}

\noindent Unless \verb`rhead` is explicitly specified, \texttt{numerica-tables}
will automatically insert a row variable header, like $x\:\!\backslash\:\negmedspace k$
in the example, where the backslash separates row from column variable.
(But see also §\ref{subsec:diagbox,-slashbox} below.) \emph{HMF}
contains a multitude of instances of this style; see Tables~9.7,
17.5, 21.1, 24.3, 27.4, etc. for examples. As noted, you can override
the default assignment to \verb`rhead` with your own specification,
but for \verb`chstyle > 0` this is no longer true.\index{rhead@\texttt{rhead}}

\subsubsection{\texttt{diagbox}, \texttt{slashbox}}

\label{subsec:diagbox,-slashbox}Rather than having $x\:\!\backslash\:\negmedspace k$
or something similar automatically inserted, you might be tempted
to use \verb`\diagbox`\index{diagbox@\texttt{\textbackslash diagbox}}
from the \texttt{diagbox} package, or \verb`\backslashbox`\index{backslashbox@\texttt{\textbackslash backslashbox}}
from the earlier \texttt{slashbox}\index{slashbox package@\texttt{slashbox} package}
package, say \verb`rhead=\diagbox{x}{k}`. Because \verb`rhead` wraps
its contents in math delimiters (\verb`$` signs) this causes a \LaTeX{}
error. The solution is to \emph{add} \verb`$` signs: \verb`rhead=$\diagbox{x}{k}$`
or, since $x$ and $k$ should be in math-italic, \verb`rhead=$\diagbox{$x$}{$k$}$`.\index{rhead@\texttt{rhead}}
You might therefore try
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},ralign=c,
                cvar=k,cstep=2,cstop=9,
                rhead=$\diagbox{$x$}{$k$}$]
      { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
but the diagonal box is far too large. One solution is to use the
optional first argument of the \verb`\diagbox` command. In that argument
one can specify an explicit \verb`height` and \verb`width` but I
prefer to work solely with the settings available in \texttt{numerica-tables}.
In that spirit, put \verb`rhsize=-4`, its smallest value. Alas this
also shrinks the displayed size of the variables $x$ and $k$. To
counteract that, explicitly specify \verb`\small$x$` and \verb`\small$k$`
for the mandatory \verb`\diagbox` arguments, the \verb`\small` overriding
the \verb`rhsize` setting to give a presentable result: 
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},ralign=c,rhsize=-4,cspec={k,2,4},
                rhead=$\diagbox{\small$x$}{\small$k$}$]
      { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$    \tabulate[rspec={x,0.2,5},ralign=c,rhsize=-4,cspec={k,2,4},
                rhead=$\diagbox{\small$x$}{\small$k$}$]
      { \sin kx }[k=3,x=0.2][4*]

However, the \texttt{booktabs} manual admonishes against the use of
vertical rules and double rules and generally advises to use as few
rules as possible. By adding a line (admittedly \emph{diagonal}, so
only half sinning) the \texttt{diagbox} or \texttt{slashbox} packages
are working in the opposite direction, more suited to tables in the
older `grid' style. In the present document I shall use $x\:\!\backslash\:\negmedspace k$
and make no further use of either \verb`\diagbox` or \verb`\backslashbox`.

\subsubsection{Built-in styles: \texttt{chstyle}}

There\index{chstyle@\texttt{chstyle}|(} are some built-in header
styles, depending on the value of the setting \verb`chstyle`. No
value for this setting corresponds to \verb`chstyle=0`, which results
in column variable values at the heads of function-value columns and
something like $x\:\!\backslash\:\negmedspace k$ at the head of the
row variable column, as previously described.

\verb`chstyle=1` changes the header of the \emph{first} function-value
column to the form \emph{variable=value}, e.g. $k=3$ in the example
below. This may be apt when a small rounding value is being used and
the resulting columns are narrow. I can find only one real instance
in \emph{HMF}, Table~26.7. Note that the row variable header, inserted
automatically, simplifies in this case to $x$ rather than $x\:\!\backslash\:\negmedspace k$. 
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},cspec={k,2,3},chstyle=1]
      { \sin kx }[k=3,x=0.2][3*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},cspec={k,2,3},chstyle=1]
      { \sin kx }[k=3,x=0.2][3*]\medskip{}

\verb`chstyle=2` changes the header of all function-value columns
to the form \emph{variable=value}.\emph{ }In \emph{HMF} examples are
Tables~7.4, 7.9, 10.10, 16.6, etc. Again, the row variable header
is inserted automatically and displays the row variable.
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},cspec={k,2,3},chstyle=2]
      { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},cspec={k,2,3},chstyle=2]
      { \sin kx }[k=3,x=0.2][4*]\medskip{}

\verb`chstyle=3` fills each column variable header with the expression
being tabulated but with the column variable replaced by its respective
values. See \emph{HMF} Tables~5.4, 8.1, 9.1, 19.1, etc. for examples. 
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},cspec={k,2,3},chstyle=3]
      { \sin kx }[k=1,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},cspec={k,2,3},chstyle=3]
      { \sin kx }[k=1,x=0.2][4*]\medskip{}

Notable here is that the column variable takes the value $1$, and
the $1$ is displayed ($\sin1x$) where it would normally be suppressed.
A similar situation can arise when the column variable takes the value
$0$. The legacy setting \verb`chstyle=4` will, in the present instance,
display $\sin1x$ as $\sin x$ but is deprecated \textendash{} given
the variety of ways that $1$ (and $0$) can appear in formulas, it
seems more straightforward in to set \verb`chstyle=0` (or omit it)
and \ldots{}

\subsubsection{User-defined header: \texttt{chead}}

\ldots{} form \verb`chead`\index{chead@\texttt{chead}} oneself, with
the relevant math delimiters (\verb`$`) and tabbing ampersands (\verb`&`).
(Always, to use \verb`chead`, \verb`chstyle` must be omitted or
set to zero.) In the example,
\begin{verbatim}
    chead=$\sin x$&$\sin3x$&$\sin5x$
\end{verbatim}
would do. It is a header for the function-value columns; a header
for the row variable column is not included. Note that there is also
no end of row mark which is inserted automatically. 

Another way to tackle the particular issue in the present case is
to form a multi-function table (see §\ref{sec:Multiple-function-tables})
of, say, $\sin x$, $\sin3x$, $\sin5x$.

\subsubsection{Alignment: \texttt{calign}}

\label{subsec:Alignment:-calign}\index{calign@\texttt{calign}}By
default the function-value columns are aligned right, \verb`calign=r`.
Also available are \verb`calign=c` for centred alignment and \verb`calign=l`
(lowercase L) for left alignment. A table containing both positive
and negative values is best set with right alignment and number padding
(\verb`*` in the trailing optional argument) to avoid the misalignment
of digits in columns caused by the presence of minus signs. (Handling
signs in tables is discussed later; see §\ref{subsec:Signs}.) Varying
numbers of digits \emph{before} the decimal point may be best handled
with the t-notation of §\ref{subsec:t-option}. Examples where a centred
alignment is helpful, centring the header above uniform columns of
figures (same number of digits before and after the decimal point)
have already been given; see §\ref{subsec:rfracFraction-form-rvar}
and §\ref{subsec:RValignment:ralign}.

\subsubsection{Nudging header entries: \texttt{chnudge}}

\label{subsec:Nudgingtheheaders:-chnudge}\index{chnudge@\texttt{chnudge}|(}As
with the row variable header, it is possible to nudge the column headers
to left or right. A positive nudge value shifts the header in the
\emph{opposite} sense to the alignment by the specified number of
mu (math units; 18 to a quad), to the left in a right alignment and
to the right otherwise. In version 3.2 of \texttt{numerica-tables,
t}he key \verb`chnudge` is a comma list of nudge values (as against
a single value in earlier versions). As with \verb`rhnudge`, only
numbers should be supplied; \texttt{numerica-tables} inserts the unit
(\verb`mu`). Often a single value is all that is required. For instance,
in the next example \verb`chnudge=9` suffices to nudge the column
headers to the left but leave the function values (with their potentially
awkward minus signs) right aligned.
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},ralign=c,
              cspec={k,2,3},chstyle=2,chnudge=9]
      { \sin kx }[k=3,x=0.2][*]
\end{verbatim}
$\Longrightarrow$     \tabulate
      [rspec={x,0.2,5},ralign=c,
         cspec={k,2,3},chstyle=2,chnudge=9]
      { \sin kx }[k=3,x=0.2][*]\medskip{}

But if the values of $k$ in this example had different numbers of
digits, no single \verb`chnudge`\index{chnudge@\texttt{chnudge}!comma list of values}
value would do. In that case give \verb`chnudge` a comma-list of
values, potentially one for each function-value column. If the comma
list has fewer entries than the number of columns, the last entry
is used for all `left over' columns. For instance, in the following
example, \verb`chnudge` is given three values, the third of which
is also used for the fourth function-value column:
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},ralign=c,cspec={k,50,3},
              chstyle=2,chnudge={13.5,9,4.5}]
      { \cos kx }[k=1,x=0.2][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},ralign=c,cspec={k,50,4},
              chstyle=2,chnudge={13.5,9,4.5}]
      { \cos kx }[k=1,x=0.2][*]\medskip{}

\emph{Negative} nudges can be useful when a column header is longer
than the displayed function values. Just for the pleasure of providing
an instance where a \emph{zero} step value\index{cstep@\texttt{cstep}!zero value}
is useful, the following small table allows an immediate comparison
of a negative nudge with no nudge. I have adjusted the separation
of the columns with \verb`\tabcolsep`\index{tabcolsep@\texttt{\textbackslash tabcolsep}}
to provide a more adequate separation of the headers:
\begin{verbatim}
    \setlength\tabcolsep{9pt}
    \tabulate[rspec={n,100,5},ralign=c,rround=0,
              cspec={k,0,2},chstyle=3,chnudge={-21,0}]
      { (1+k/n)^n }[k=1,n=100][3*]
    \setlength\tabcolsep{6pt}
\end{verbatim}
$\Longrightarrow$     \setlength\tabcolsep{9pt}
    \tabulate[rspec={n,100,5},ralign=c,rround=0,
              cspec={k,0,2},chstyle=3,chnudge={-21,0}]
      { (1+k/n)^n }[k=1,n=100][3*]
    \setlength\tabcolsep{6pt}\index{chnudge@\texttt{chnudge}|)}

\subsubsection{Math style: \texttt{chmath}}

The \verb`chmath` setting is a comma list of the values \verb`s`,
\verb`t`, or \verb`d` presenting the header of the corresponding
function-value column in either scriptstyle, textstyle or displaystyle
respectively. The setting is initialized to \verb`t`. If there are
more function-value columns in the table than entries in \verb`chmath`
the last value in \verb`chmath` is used for the extra columns. See
the first table in §\ref{sec:Multiple-function-tables} (on multi-function
tables) for an example of this setting's use.

\subsubsection{Font size: \texttt{chsize}}

\label{subsec:chsize}The font size of the header row entries can
be changed with the setting \verb`chsize`\index{chsize@\texttt{chsize}}
which is a comma list of integer values, each integer in the range
$-4$ to $5$ corresponding to a \LaTeX{} font size command according
to the scheme

\begin{tabular}{rlcrl}
$-4$ & \verb`\tiny` &  & $1$ & \verb`\large`\tabularnewline
$-3$ & \verb`\scriptsize` &  & $2$ & \verb`\Large`\tabularnewline
$-2$ & \verb`\footnotesize` &  & $3$ & \verb`\LARGE`\tabularnewline
$-1$ & \verb`\small` &  & $4$ & \verb`\huge`\tabularnewline
$0$ & \verb`\normalsize` &  & $5$ & \verb`\Huge`\tabularnewline
\end{tabular}\medskip{}

\noindent Should there be more columns in the table than entries in
\verb`chsize`, then the last value in \verb`chsize` is used for
the extra columns. In the following example, the font size has been
tweaked by means of the setting \verb`chsize={1,-1,1}` to bring the
character size of the column headings into greater conformity:
\begin{verbatim}
     \tabulate[ff,rspec={x,0.5,4},chsize={1,-1,1}]
       { \frac{ e^{2x}-1}{ e^{2x}+1},\tanh x,
         \frac{ \sinh x}{\cosh x}} [x=0.5] [4*]
\end{verbatim}
$\Longrightarrow$          \tabulate[ff,rspec={x,0.5,4},chsize={1,-1,1}]
       { \frac{ e^{2x}-1}{ e^{2x}+1}, \tanh x ,
         \frac{ \sinh x}{\cosh x}} [x=0.5] [4*]\index{column variable header!formatting|)}\index{column variable header!multi-column|)}

\section{Multiple functions in a single table}

\label{sec:Multiple-function-tables}As\index{multi-function tables|(}
noted in §\ref{subsec:Adjoining-tables}, using adjoining tables in
order to tabulate more than one function at a time is `clunky'.
Much better is to enter the functions in the main argument of a \verb`\tabulate`
command separated by a specified mark, then alert \verb`\tabulate`
that this has happened by setting \texttt{numerica}'s \verb`ff`\index{ff@\texttt{ff}}
key. 

The default multi-function delimiter\index{multi-function tables!function delimiter}
is the comma if the decimal point is a dot (period), or the semicolon\index{decimal comma!semicolon use}
if it is a comma (\texttt{numerica} loaded with the \verb`comma`
package option\index{comma package option@\texttt{comma} package option}).
With the default delimiter  it suffices to enter \verb`ff` in the
settings option; otherwise enter \verb`ff=<mark>` there. For example
\verb`ff=|` makes the `pipe' character the multi-formula delimiter.
If the \verb`ff` key is overlooked then multiple formulas in the
main argument will almost certainly cause a \LaTeX{} error.

The following example (of the so-called `Einstein functions') illustrates
both a multi-function table using the default comma as separator,
and the use of \verb`chmath`\index{chmath@\texttt{chmath}} to shrink
the function in the last column. If not shrunk, $\ln(1-e^{-x})$ appears
too large compared to the symbols in the fractions in the middle columns:
\begin{verbatim}
    \tabulate[ff,rspec={x,0.3,4},rround=2,rhnudge=9,
              chnudge={0,9,0},chmath={t,t,s}]
      { \frac{x^2e^x}{(e^x-1)^2}, \frac{x}{e^x-1},
          \ln(1-e^{-x}) }[x=0.15][5*]
\end{verbatim}
$\Longrightarrow$     \tabulate[ff,rspec={x,0.3,4},rround=2,rhnudge=9,
              chnudge={0,9,0},chmath={t,t,s}]
      { \frac{x^2e^x}{(e^x-1)^2}, \frac{x}{e^x-1},
          \ln(1-e^{-x}) }[x=0.15][5*]     \medskip{}

In the next example I transpose this table, not because it is a good
example of a table that might benefit from this, but because it illustrates
a problem that can arise when a multi-function table is transposed.
On transposition the current header row becomes the row variable column:
\index{transpose@\texttt{transpose}!examples of use}
\begin{verbatim}
    \tabulate[ff,rspec={x,0.3,4},rround=2,ralign=l,
              chnudge={0,9},rules=hB,chstyle=1,transpose]
      { \frac{x^2e^x}{(e^x-1)^2}, \frac{x}{e^x-1},
           \ln(1-e^{-x}) }[x=0.15][5*]
\end{verbatim}
$\Longrightarrow$     \tabulate[ff,rspec={x,0.3,4},rround=2,ralign=l,
              chnudge={0,9},rules=hB,chstyle=1,transpose]
      { \frac{x^2e^x}{(e^x-1)^2}, \frac{x}{e^x-1},
          \ln(1-e^{-x}) }[x=0.15][5*]   \medskip{}

\noindent What is immediately noticeable is the disparity in size
of the characters forming $\ln(1-e^{-x})$ compared to those in the
fractions above it. This was the reason for the \verb`chmath={t,t,s}`
setting in the original table, but \verb`rmath`\index{rmath@\texttt{rmath}}
is a single value applying to the \emph{whole} row variable column,
so that solution is not available. Two possibilities come to mind.
One  is to force the function being tabulated into scriptstyle by
tabulating
\begin{verbatim}
    \vphantom{\frac11}\scriptstyle\ln(1-e^{-x})
\end{verbatim}
(The phantom ensures the spacing between the rows is the same.) This
new function produces exactly the same values as the previous one
did. The other possibility is to set \verb`rmath=d`, forcing all
members of the row variable column, in particular the fractions, into
displaystyle. With this setting the \verb`\scriptstyle` can be omitted
(but the phantom retained).

Another example, now using \verb`ff=|` as the function separator,
places the row variable column on both sides of the table (\verb`rpos=3`)\index{rpos@\texttt{rpos}}
and uses the \verb`o` setting (see \texttt{numerica.pdf}) to indicate
that arguments of the trig functions are in degrees:
\begin{verbatim}
    \tabulate[ff=|,o,rpos=3,rround=0,chnudge=9,
              rvar=\theta,rstep=15,rstop=90]
      { \sin \theta | \cos \theta }[\theta=0][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[ff=|,o,rpos=3,rround=0,chnudge=9,
              rvar=\theta,rstep=15,rstop=90]
      { \sin \theta |\cos \theta }[\theta=0][*]\medskip{}

The table suggests a space saving possibility: since $\sin$ and $\cos$
are complementary functions\index{complementary functions} ($\sin(90-\theta)=\cos\theta$),
values in the bottom half of the table duplicate values in the top
half, only with the columns reversed. This is the reason for the \verb`rpos=4`\index{rpos@\texttt{rpos}}
setting discussed in §\ref{subsec:Second-row-var-col}, which enables
complementary functions to be tabulated in `half tables'.\index{multi-function tables|)}

\section{Whole-of-table formatting}

\label{sec:Whole-of-table-formatting}
\begin{table}[t]
\centering
\centering{}\caption{\label{tab:Table-formatting-settings}\label{tab:Table-formatting}Table
formatting}
\begin{center}
\begin{tabular}{ll>{\raggedright}p{4cm}l}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small initial}\tabularnewline
\midrule
{\small\texttt{ctitle}} & {\small token(s)} & {\small collective title for columns} & \tabularnewline
{\small\texttt{csubttl}} & {\small token(s)} & {\small subtitle row for} {\small function- value cols} & \tabularnewline
{\small\texttt{calign}} & {\small char(}{\small\texttt{r}}{\small /}{\small\texttt{c}}{\small /}{\small\texttt{l}}{\small )} & {\small column alignment} & {\small\texttt{r}}\tabularnewline
{\small\texttt{headless}} &  & {\small suppress header row} & \tabularnewline
{\small\texttt{foot}} & {\small token(s)} & {\small table-wide footer row} & \tabularnewline
{\small\texttt{rules}} & {\small char(s)} & {\small horizontal rule spec.} & {\small\texttt{ThB}}\tabularnewline
{\small\texttt{norules}} &  & {\small cancel all rules} & \tabularnewline
{\small\texttt{rpos}} & {\small int (}{\small\texttt{0}}{\small\ldots}{\small\texttt{4}}{\small )} & {\small row variable col. position(s)} & {\small\texttt{1}}\tabularnewline
{\small\texttt{rbloc}} & {\small integer comma list} & {\small row block specification} & \tabularnewline
{\small\texttt{rblocsep}} & {\small length} & {\small extra space between row blocks} & {\small\texttt{1ex}}\tabularnewline
{\small\texttt{transpose}} &  & {\small show funct. vals in rows} & \tabularnewline
{\small\texttt{valign}} & {\small char (}{\small\texttt{t}}/{\small\texttt{m}}/{\small\texttt{b}}{\small )} & {\small vertical alignment of table relative to text baseline} & {\small\texttt{m}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}
There\index{formatting!title, subtitle, footer|(} are a number of
settings that affect the appearance of the table as a whole, things
like the position of the row variable column, the grouping of function
values in a column into blocks to aid readability, the presence of
horizontal rules, the use of a collective column title, or of a footer
row, or display of the table `horizontally' in rows rather than
columns. I discuss these here (Table~\ref{tab:Table-formatting}).

\subsection{Title for function-value columns: \texttt{ctitle}}

\label{subsec:Title:-ctitle-setting}\index{ctitle@\texttt{ctitle}|(}The\index{title row|(}
function-value columns have individual headings, formatted in the
various ways provided by the settings already discussed, but it can
also be helpful to have a collective title for these columns. For
instance our table of Einstein functions in §\ref{sec:Multiple-function-tables}
would have benefited from this. The need is met with the \verb`ctitle`
key. This can be set to whatever you like (e.g. \verb`ctitle=Fred`),
taking care to shield commas with braces. By default, the title is
set between math delimiters (\verb`$` signs). For that reason in
the example I have placed it inside a \verb`\text` command: 
\begin{verbatim}
    \tabulate[ff,rspec={x,0.15,6},rround=2,rhnudge=9,
              chnudge={0,9,0},chmath={t,t,s},
              ctitle=\text{Einstein functions}]
      { \frac{x^2e^x}{(e^x-1)^2}, \frac{x}{e^x-1},
          \ln(1-e^{-x}) }[x=0.15][5*]
\end{verbatim}
$\Longrightarrow$     \tabulate[ff,rspec={x,.15,6},rround=2,rhnudge=9,
              chnudge={0,9,0},chmath={t,t,s},
              ctitle=\text{Einstein functions}]
      { \frac{x^2e^x}{(e^x-1)^2},
          \frac{x}{e^x-1},\ln(1-e^{-x}) }[x=0.15][5*]      

\medskip{}

There are two built-in values for the \verb`ctitle` key: \verb`ctitle=*`,\index{ctitle@\texttt{ctitle}!{*},{*}{*} settings}
which forms the title from the function being tabulated, and \verb`ctitle=**`
which uses the function and vv-list for the title. Obviously these,
particularly the latter, could easily become too long to be useful.
An example of \verb`ctitle=**` is presented later in §\ref{subsec:Rules:-rules-setting},
but inclusion of the vv-list in the title in the next example would
be pointless since the row and column variables of the table are the
only members of the vv-list; \verb`ctitle=*` suffices:
\begin{verbatim}
    \tabulate[rspec={n,1,5},rround=0,
              cspec={m,1,4},chstyle=2,ctitle=*]
      { \cos(m\pi/n) }[n=3,m=2][*4]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={n,1,5},rround=0,
              cspec={m,1,4},chstyle=2,ctitle=*]
      { \cos(m\pi/n) }[n=3,m=2][*4]

\subsection{Between header \& title: \texttt{csubttl}}

\label{subsec:subtitle-row}Some\index{csubttl@\texttt{csubttl}}\index{title row!subtitle row}
tables need more header or title material than can be comfortably
accommodated in either row alone. For examples, see \emph{HMF} Tables
7.9, 17.7, 21.1, and 26.7. One way of handling this problem is to
resort to more complicated environments in header and title rows.
Another, more direct way, is to insert a row between the header and
title rows by means of the key \verb`csubttl`, a contraction of `c(olumn
variable) subtitle'. (In version 2 of \texttt{numerica-tables} the
name \verb`cmidrow`\index{cmidrow@\texttt{cmidrow }|see{\texttt{csubttl}}}
was used; that is still available, but deprecated.) Like \verb`chead`
and \verb`ctitle`, \verb`csubttl` is limited to the span of the
column variable (or function-value) columns only. The content of \verb`csubttl`
is entirely up to the user, including insertion of sufficient \verb`&`
characters and math delimiters (if required).

An example where the subtitle row serves an explanatory role is shown
in the next table. I have used \verb`\sfrac` from the \texttt{xfrac}
package for neater fractions:
\begin{verbatim}
     \tabulate[ff,rspec={x,1,6,0},chnudge={22.5,18},
      ctitle=\text{Hyperbolic functions},
      csubttl=\multicolumn{3}{c}{\small$\sinh x=\sfrac12
        (e^x-e^{-x}),\ \cosh x=\sfrac12(e^x+e^{-x})$}]
       { \sfrac12\,e^x, \sinh x, \cosh x } [{x}=0][8*]       
\end{verbatim}
$\Longrightarrow$     \tabulate[ff,rspec={x,1,6,0},chnudge={22.5,18},
      ctitle=\text{Hyperbolic functions},
      csubttl=\multicolumn{3}{c}{\small$\sinh x=\sfrac12
        (e^x-e^{-x}),\ \cosh x=\sfrac12(e^x+e^{-x})$}]
       { \sfrac12\,e^x, \sinh x, \cosh x } [{x}=0][8*]\index{ctitle@\texttt{ctitle}|)}\index{title row|)}
\begin{verbatim}

\end{verbatim}

\subsection{Suppress/show header row}

\label{subsec:Header-row-suppress}By default\index{headless@\texttt{headless}}
the header row in a table is shown. It carries essential information
as to the table's contents, but there are occasions when it should
be suppressed. An example occurs in §\ref{subsec:rdata-rfile-rverb}
where a table listing fractions of $\pi$ and their values is shown.
The header there serves no purpose. To suppress it, use the \verb`headless`\footnote{There was a mix-up in the documentation of version 3.1 of \texttt{numerica-tables}
between what was entered in the table of whole-of-table settings and
what was described in this section.} setting. \index{norules@\texttt{norules}}
\begin{verbatim}
    \def\mydata{\sfrac14\,\pi,\sfrac13\,\pi,\sfrac12\,\pi,
                  \sfrac23\,\pi,\sfrac34\,\pi,\pi}
    \tabulate[rdata=\mydata,rverb=1,rvar=k,headless,
                norules]{ k }[*]
\end{verbatim}
$\Longrightarrow$     \def\mydata{\sfrac14\,\pi,\sfrac13\,\pi,\sfrac12\,\pi,
                  \sfrac23\,\pi,\sfrac34\,\pi,\pi}
    \tabulate[rdata=\mydata,rverb=1,rvar=k, headless,
                norules]{ k }[*]

\subsection{Footer row: \texttt{foot}}

\label{subsec:Footer-row:-foot}Some\index{foot@\texttt{foot}}\index{footer row}
tables have a footer row and \verb`numerica-tables` allows such a
row to be inserted, but its entire content, with two exceptions, is
the responsibility of the user, including insertion of the necessary
number of tab characters \verb`&`. This will usually be $1$ less
than the total number of columns (including row variable columns)
in the table \textendash{} or some adjustment thereof if you use \verb`\multicolumn`.
(\emph{HMF} uses the footer mainly for cryptic descriptions of the
accuracy and needs of interpolation methods.) 

You can put into the footer what you wish with the setting \verb`foot=<tokens>`.

There are two exceptions: 
\begin{itemize}
\item when \verb`foot=''`\index{foot@\texttt{foot}} (two single quote
marks), suggesting ditto marks, the footer row is set equal to the
header row with one change, the backslash in row variable headers
of the form $n\:\!\backslash\:\!m$ are converted to forward slashes,
e.g. $n\:\!/\:\!m$; 
\item when \verb`foot=*`\index{foot@\texttt{foot}}, the footer is set
equal to the header in \emph{reversed} order \textendash{} the last
item first, and so on. This is useful for tabulating complementary
functions like sine and cosine or, more generally, $f(x)$ and $g(x)$
where $g(x)=f(k-x)$ for some constant $k$. Values for the complementary
function are read from the bottom up and require a reversed row variable
column on the right of the table; see the third example in §\ref{subsec:Second-row-var-col}.
\end{itemize}

\paragraph{Footer functions}

In\index{footer functions} versions of \texttt{numerica-tables} before
version 3 it was possible to perform certain simple operations on
columns \textendash{} calculate the sum, the average and maximum and
minimum values. This is no longer so in version 3. Not only does it
seem tangential to the primary function of the \verb`\tabulate` command
but it was also acutely dependent on the format of the numbers being
operated on \textendash{} a change in the number-format option could
cause a \LaTeX{} error.\index{formatting!title, subtitle, footer|)}

\subsection{Horizontal rules: \texttt{rules}}

\label{subsec:Rules:-rules-setting}\index{rules@\texttt{rules}|(}\index{formatting!rules|(}The
\texttt{booktabs} \index{booktabs@\texttt{booktabs}} package which
\texttt{numerica-tables} uses is most emphatic that one should `1.
Never, ever use vertical rules.\index{rules@\texttt{rules}!edicts}
2. Never use double rules.' Most of the tables proper in \emph{HMF}
lack rules of any kind although closer inspection shows smaller tables
within the text generally \emph{are} delimited by horizontal rules
(often also with vertical rules). In the various examples in the present
document I have used horizontal rules because these too are tables
within text. Some form of delineation seems necessary although I think
I have `over ruled'. The question should always be: is a rule necessary
at all? Usually, less is more. (Although many of \emph{HMF}'s tables
are inelegantly typeset, I have used it as a valuable resource for
the variety of structures needed to present a multitude of different
kinds of numerical data.) 

The \verb`rules` key allows one to specify precisely which rules
are used. The content of the key is a `word' \textendash{} a sequence
of letters \textendash{} where the characters have the significance
and default thicknesses \textendash{} from the \texttt{booktabs}\index{booktabs@\texttt{booktabs}}
package \textendash{} shown in Table~\ref{tab:Rules}. (To adjust
the thickness of rules used, see §\ref{subsec:Trim-and-thickness}
below.) No particular order is required in the \verb`rules` specification;
\verb`rules=hBT` would work as well as \verb`rules=ThB` but the
latter, reflecting the order of the rules in the table, is simpler
to grasp at a glance.

In previous versions of \texttt{numerica-tables} the initial value
of this key was \verb`rules=ThB` so that if the \verb`rules` key
was not explicitly set then a rule was automatically drawn at the
top of the table, at the bottom of the table, and beneath the header
row. From version 3.2, no initial value is assigned to \verb`rules`.
Instead, a package option of the same name\index{rules@\texttt{rules}!package option}\index{package options}
can be set (see §\ref{chap:Introduction}) which is initialized to
\verb`rules=ThB` to maintain previous behaviour. What is gained is
that by assigning a different value to the package option it is now
possible for a user to alter the default behaviour. A \verb`norules`
package option is also available if you want your tables free of rules
by default. The default set by the package option can always be overridden
for an individual table by means of the \verb`rules` \emph{setting}.

\subsubsection{Header and footer rules}

The header (resp. footer) rule specified by including \verb`h` (resp.
\verb`f`) in the \verb`rules` specification spans the whole table,
from the left of the first column to the right of the last column.
It can give a better visual appearance sometimes to trim (shorten)
these rules both left and right. This is specified by priming the
\verb`h` (resp. \verb`f`) in the specification. Thus \verb`rules=hf`
will produce header and footer rules spanning the whole table; \verb`rules=h'f'`
will produce slightly shorter header and footer rules trimmed by the
default $0.5\text{\,em}$ at both ends \textendash{} unless some other
trim value has been specified; see §\ref{subsec:Trim-and-thickness}
below.

\subsubsection{Title and subtitle rules}

\begin{table}[t]
\centering
\centering{}\caption{\label{tab:Rules}Rules. {\small (In the \textquoteleft span\textquoteright{}
column, \textquoteleft fv\textquoteright =function-value; \textquoteleft rv\textquoteright =row
variable.)}}
\begin{center}
\begin{tabular}{lll>{\raggedright}p{2.2cm}V{\linewidth}V{\linewidth}}
\toprule 
{\footnotesize char} & {\footnotesize rule} & {\footnotesize position} & {\footnotesize span} & {\footnotesize trim} & {\footnotesize default rule thickness}\tabularnewline
\midrule
{\small\texttt{T}} & {\footnotesize top} & {\footnotesize above table} & {\footnotesize table} &  & {\footnotesize\texttt{\textbackslash heavyrulewidth=.08em}}\tabularnewline
{\small\texttt{t}} & {\footnotesize title} & {\footnotesize below title} & {\footnotesize fv cols} & {\small\texttt{.5em}} & {\footnotesize\texttt{\textbackslash cmidrulewidth=.03em}}\tabularnewline
{\small\texttt{s}} & {\footnotesize subtitle} & {\footnotesize below subtitle} & {\footnotesize fv cols ($1$ rv col.) table ($2$ rv cols)} & {\small\texttt{.5em}}{\small\par}

{\small\texttt{.5em}} & {\footnotesize\texttt{\textbackslash cmidrulewidth=.03em}}{\footnotesize\par}

{\footnotesize\texttt{\textbackslash lightrulewidth=.05em}}\tabularnewline
{\small\texttt{h}} & {\footnotesize header} & {\footnotesize below header} & {\footnotesize table} &  & {\footnotesize\texttt{\textbackslash lightrulewidth=.05em}}\tabularnewline
{\small\texttt{h'}} &  &  &  & {\small\texttt{.5em}} & {\footnotesize\texttt{\textbackslash cmidrulewidth=.03em}}\tabularnewline
{\small\texttt{f}} & {\footnotesize footer} & {\footnotesize above footer} & {\footnotesize table} &  & {\footnotesize\texttt{\textbackslash lightrulewidth=.05em}}\tabularnewline
{\small\texttt{f}}' &  &  &  & {\small\texttt{.5em}} & {\footnotesize\texttt{\textbackslash cmidrulewidth=.03em}}\tabularnewline
{\small\texttt{B}} & {\footnotesize bottom} & {\footnotesize below table} & {\footnotesize table} &  & {\footnotesize\texttt{\textbackslash heavyrulewidth=.08em}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}
If \verb`t` is included in the \verb`rules` specification, then
a rule will be drawn beneath the title. This will span the function-value
columns only and be trimmed at each end (by default by 0.5 em). If
you are also using a subtitle row, between title and header rows,
and want a rule beneath that too, then include \verb`s` in the setting
\textendash{} for instance \verb`rules=TtshB`. (For legacy reasons,
\verb`m` \textendash{} from `midrow' \textendash{} can also be
used instead of \verb`s`.) To my eye rules beneath \emph{both} title
and subtitle don't work; a rule beneath the subtitle alone gives a
better result \textendash{} if a rule is needed at all. 

The span of the subtitle rule depends on whether the row variable
column is placed only on one side of the table or on both (\verb`rpos<2`
or \verb`rpos>2`). If on only one then the  rule spans only the function-value
columns \textendash{} like the title rule. If on both then the subtitle
rule spans the table but is trimmed by 0.5~em at each end. To the
author's eye, the extra span in this second case gives a better sense
of the row variable columns as part of the table as a whole rather
than as optional `clip ons' at the edges.

In the example table below, a rule for the column title has been specified
(the \verb`t` in the setting \verb`rules=TthB`). Also note the use
of \verb`ctitle=**`\index{ctitle@\texttt{ctitle}!{*},{*}{*} settings}.
The formula contains an extra parameter $a$, assigned a value in
the vv-list, so that it now makes sense to display the vv-list in
the column title (the braces around \verb`k` and \verb`x` in the
vv-list ensure they don't display).
\begin{verbatim}
    \tabulate[rspec={x,0.25,5,2},cspec={k,0.25,3,2},
              rhnudge=9,chstyle=2,ctitle=**,rules=TthB]
      { a\sin kx }[a=2/\pi,{k}=3,{x}=0.25][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.25,5,2},cspec={k,0.25,3,2},
              rhnudge=9,chstyle=2,ctitle=**,rules=TthB] 
      { a\sin kx }[a=2/\pi,{k}=3,{x}=0.25][*]

\subsubsection{Trim and thickness of rules}

\label{subsec:Trim-and-thickness}The trim applied to rules \textendash{}
the amount cut from the ends \textendash{} is the \texttt{booktabs}'\index{booktabs@\texttt{booktabs}}
default value, $0.5$~em. To change it, enter the command
\begin{verbatim}
    \setlength\cmidrulekern{<trim>}
\end{verbatim}
in the preamble where \verb`<trim>` is a length, something like \verb`0.5em`,
\verb`2.5mm`, etc. (Or, one could write, e.g. \verb`\cmidrulekern=2.5mm`.)
The new default will apply to the settings \verb`h'`, \verb`f'`,
\verb`t` and sometimes \verb`s`, since these all use \texttt{booktabs}'
\verb`\cmidrule`.

To change the thickness of a rule from its default value, enter new
values in the preamble for any or all of \verb`\heavyrulewidth`\index{heavyrulewidth@\texttt{\textbackslash heavyrulewidth}},
\verb`\lightrulewidth`\index{lightrulewidth@\texttt{\textbackslash lightrulewidth}},
\verb`\cmidrulewidth`\index{cmidrulewidth@\texttt{\textbackslash cmidrulewidth}},
e.g.
\begin{verbatim}
    \setlength\heavyrulewidth{<width>}
\end{verbatim}
where \verb`<width>` is a length. The values listed in Table~\ref{tab:Rules}
are the default values, from the \texttt{booktabs}\index{booktabs@\texttt{booktabs}}
package, used in \texttt{numerica-tables}.\index{rules@\texttt{rules}|)}\index{formatting!rules}

\subsection{Second row variable column: \texttt{rpos=3,4}}

\label{subsec:Second-row-var-col}\index{row variable column!position in table}\index{rpos@\texttt{rpos}|(}The
settings \verb`rpos=0,1,2` have been discussed earlier; {\ttfamily\verb`rpos=3`}
duplicates the row variable column on the right of the table, with
the proviso that if the header of the row variable column on the left
is automatically generated and like $x\!\:\backslash\!\:k$, the header
on the right is changed to a form like $k\!\:/\!\:x$. Repeating the
first example of §\ref{subsec:Headerstyle:-multi-column-case} but
with the setting \verb`rpos=3`, 
\begin{verbatim}
    \tabulate[rpos=3,rspec={x,0.2,5},
      ccspec={k,2,cstop=7,chnudge=18]
        { \sin kx }[k=3,x=0.2][4*]
\end{verbatim}
$\Longrightarrow$    \tabulate[rpos=3,rspec={x,0.2,5},
      cvar=k,cstep=2,cstop=7,chnudge=18]
        { \sin kx }[k=3,x=0.2][4*]\medskip{}

\noindent An example of using \verb`rpos=3` in a multi-function table
is provided near the end of §\ref{sec:Multiple-function-tables}. 

The \verb`rpos` key can also take the value $4$. \verb`rpos=4`
adds the row variable column to both left and right sides of the table,
but the values displayed in the right column are a \emph{function
of those in the left column} ({\ttfamily\verb`rpos=3`}
corresponds to the function being the identity). The value of the
key \verb`rvar'`\index{rvar@\texttt{rvar'}|(} specifies the function
used.
\begin{verbatim}
    \tabulate[rpos=4,rspec={x,0.5,5,2},rhnudge=9,
                cspec={k,2,3},chstyle=2,ctitle=*,
                  rvar'=x^2,rhnudge'=4]
        {  \sin(kx^2) }[k=3,x=1][4*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rpos=4,rspec={x,0.5,5,2},rhnudge=9,
                chstyle=2,cspec={k,2,3},ctitle=*,
                  rvar'=x^2,rhnudge'=4]
        {  \sin(kx^2) }[k=3,x=1][4*]\medskip{}

\noindent If no value is given to \verb`rvar'` (or if \verb`rvar'=*`)
then the row variable column on the right of the table is the left
column reversed (with the possible exception of its header).
\begin{itemize}
\item Note:\index{transpose@\texttt{transpose}!and rpos@and \texttt{rpos}}
\verb`rpos=4` is incompatible with the \verb`transpose` setting.
If both settings are used, the table will be transposed as if \verb`rpos=3`.
\end{itemize}
When \verb`chstyle` is not zero, as in the example above, and \verb`rhead'`
is omitted or empty, the header for the right-hand row variable column
is automatically generated, displaying the value of the \verb`rvar'`
key. The positioning of the right-hand row variable header can be
adjusted with the use of \verb`rhnudge'`\index{rhnudge@\texttt{rhnudge}'}
as in the example. If \verb`chstyle=0`, the header is up to the user
to supply by assigning a value to the key \verb`rhead'`\index{rhead@\texttt{rhead'}}.
In all cases, \verb`rhead'` will override any automatically generated
header.

The sine and cosine are complementary functions\index{complementary functions|(};
when working in degrees (the \verb`o` setting in the next example),
$\cos\theta=\sin(90-\theta)$. We can exploit this fact to halve the
table size needed to tabulate the two functions. The following `toy'
example (see\emph{ HMF} Tables 4.10\textendash 4.12 for the real thing)
of a multi-function table (hence the \verb`ff`\index{ff@\texttt{ff}}
setting) also gives an example of the use of the \verb`foot=*`\index{foot@\texttt{foot}}
setting, and the primed versions of header and footer rules\index{rules@\texttt{rules}}.
Note also the use of an  expression in the third element of \verb`rspec`.
The \verb`rhead'` setting is of no significance; it simply avoids
a large block of white space.
\begin{verbatim}
    \tabulate[ff,o,rpos=4,rules=Th'f'B,foot=*,calign=c,
               rspec={\theta,9,1+45/9,0},
                 rvar'=90-\theta,rhead'=-]
      { \sin\theta,\cos\theta }[\theta=0][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[ff,o,rpos=4,rules=Th'f'B,foot=*,calign=c,
               rspec={\theta,9,1+45/9,0},
                 rvar'=90-\theta,rhead'=-]
      { \sin\theta,\cos\theta }[\theta=0][*]\medskip{}

\noindent The values of sines from $0$ to $45$ degrees are read
downwards from the first column of function-values, and from 45 to
90 degrees are read upwards from the second column of function-values.
For cosines it is downwards from the second column and upwards from
the first column. The reversed footer line indicates the change of
columns to use.

Although there is a significant space saving with tables like this,
they are not `kind to the reader'. They require a certain concentration
to read and should be avoided unless space is seriously constrained.
\emph{HMF} Tables 6.1 and 6.2 are tables of the gamma function and
its relatives where $y=x-1$ is used in the row variable column on
the right (stemming from $x!=\Gamma(x-1)$); \emph{HMF }Table 6.5
in effect uses $\langle1/x\rangle$ (the nearest integer to $1/x$)
for the row variable on the right.\index{complementary functions|)}

\subsection{Separating blocks of rows: \texttt{rbloc}}

Readability\index{formatting!grouping rows|(}\index{row grouping|(}
of long columns of figures can be aided by adding extra white space
between blocks of rows. This is achieved with the \verb`rbloc`\index{rbloc@\texttt{rbloc}|(}
key:
\begin{verbatim}
    rbloc = <comma list of positive integers>
\end{verbatim}
specifies how many rows belong to each block. For example, \verb`rbloc={5,5,6}`
breaks the table into blocks of $5$ rows, $5$ rows, then $6$ rows.
If the number of rows in the table is greater than the sum of the
entries in the comma list, then division into blocks continues as
specified by the last entry in the comma list. Thus \verb`rbloc=5`
(strictly \verb`rbloc={5}` but the braces can be omitted in this
case since no comma is enclosed) divides a table into blocks of $5$
rows; \verb`rbloc={1,5}` divides a table into $1$ row followed by
blocks of $5$ rows. A division of this kind may be appropriate when,
say, the row variable runs from $0$ to $1$ in increments of $0.1$
\textendash{} there are $11$ rows of which the first (when the row
variable is zero) may have distinctive values. 

The dominant practice in \emph{HMF} is division into blocks of (generally)
$5$ rows, many of which start with a zero value for the row variable.
Rather than isolate this initial value, they include it in the first
block of $5$, then continue with blocks of $5$ until a single isolated
row is left at the bottom of the page or the table. There seems to
be a psychological need to finish a page or table with the row variable
set to a nice round number. Thus: tabulate from $0$ to $10$ rather
than $0$ to $9$, from $0$ to $1$ rather than $0$ to $0.9$, and
even from $0$ to $30$ or $0$ to $2$ rather than $0$ to $29$
or $0$ to $1.9$. Using blocks of $5$ the consequence is that there
is always an isolated line at the end \textendash{} a kind of punctuation
mark to signal the end of the page or the table.

In the next example I have divided the rows into blocks of $4$ but
separated the special values of the first and last rows by setting
\verb`rbloc={1,4}`.
\begin{verbatim}
    \tabulate[ff,o,rspec={\theta,10,1+90/10,0},rbloc={1,4}]
        { \sin\theta, \cos\theta }[\theta=0][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[ff,o,rspec={\theta,10,1+90/10,0},rbloc={1,4}]
        { \sin\theta, \cos\theta }[\theta=0][*]\index{rbloc@\texttt{rbloc}|)}

\subsubsection{Adjusting the extra space\texttt{ rblocsep} }

By default \texttt{numerica-tables} sets the extra space between blocks
of rows at \verb`1 ex`. This value can easily by changed with the
setting \verb`rblocsep=<length>`\index{rblocsep@\texttt{rblocsep}}.
The units need to be included in the specification. \index{formatting!grouping rows|)}\index{row grouping|)}

\subsection{\textquoteleft Horizontal\textquoteright{} tables: \texttt{transpose}}

\label{subsec:Transpose}Traditionally\index{tables!`horizontal'/`vertical'|(}\index{transpose@\texttt{transpose}|(}
in a book of mathematical tables the function values are arranged
vertically in columns. This is why the row variable has been given
primacy in \texttt{numerica-tables}. But that may not always be what
is wanted nor make the best use of available space. From version 3.2.0
of \texttt{numerica-tables} there is a setting, \verb`transpose`,
that will convert a `vertical' table (function values in columns)
to a `horizontal' table (function values in rows). The table is
\emph{constructed} as if the function values will be listed vertically
in columns, but with \verb`transpose` entered in the settings option
the table is displayed with the function values listed in rows. 

(Strictly, \verb`transpose` `is really' \verb`transpose=1`, but
\verb`transpose` alone suffices; \verb`transpose=0` corresponds
to the default situation of function values in columns.) 

With transposition, the row variable column and header row are interchanged.
The building of the row variable column and header row occurs \emph{before}
transposition. Hence \verb`rround` and \verb`chround` are applied
to these elements \emph{before} transposition. It is the rounded values
of the row and column variables that are used to calculate function
values. But `cosmetic effects' like placement of the (new) row variable
column or (new) header row, as well as header styling (\verb`chstyle`),
alignment (\verb`ralign`, \verb`calign`), nudging (\verb`rhnudge`,
\verb`chnudge`), fonts (\verb`rfont`, \verb`chfont`), and choice
of display-, text- or scriptstyle (\verb`rmath`, \verb`chmath`)
are applied \emph{after} transposition. A footer row is created \emph{after}
transposition.
\begin{itemize}
\item Transposition\index{transpose@\texttt{transpose}!and rpos@and \texttt{rpos}}
does not work for tables using a second row variable column with \verb`rpos=4`.
Such a table \emph{is} transposed, but treated as if \verb`rpos=3`.
\end{itemize}
In a `shallow' transposed table \textendash{} like one with a single
row of function values \textendash{} rules should be used sparingly
if at all. See the example at §\ref{sec:Row-variable-settings} where
the \verb`norules` setting is used. In the following example, with
three function-value rows resulting from the transposition, a single
header rule suffices. 
\begin{verbatim}
    \tabulate[rspec={x,0.2,5},ralign=c,cspec={k,0.25,3,2},
                chstyle=1,rules=h',transpose]
      { \sin kx }[k=3,x=0.2][*]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,5},ralign=c,cspec={k,0.25,3,2},
                chstyle=1,rules=h',transpose]
      { \sin kx }[k=3,x=0.2][*]\index{tables!`horizontal'/`vertical'|)}\index{transpose@\texttt{transpose}|)}

\subsection{Table placement}

\label{sec:Table-placement}Tables\index{tables!position on line/page}
can be nudged vertically with the \LaTeX{} commands \verb`\bigskip`\index{bigskip@\texttt{\textbackslash bigskip}},
\verb`\medskip`, \verb`\smallskip`\index{smallskip@\texttt{\textbackslash smallskip}},
usually about $1$, $\sfrac12$ and $\sfrac14$ line spaces (with
stretch and shrink). \texttt{booktabs}\index{booktabs@\texttt{booktabs}}
provides \verb`\abovetopsep`\index{abovetopsep@\texttt{\textbackslash abovetopsep}}
and \verb`\belowbottomsep`\index{belowbottomsep@\texttt{\textbackslash belowbottomsep}},
both set by default to \verb`0ex` and easily changed by writing,
e.g. \verb`\setlength\abovetopsep{1.25ex}` (or \verb`\abovetopsep=1.25ex`)
if you want to insert \verb`1.25ex` of space above the table (perhaps
to fit captions). 

\subsubsection{Vertical~alignment}

\label{subsec:Verticalalignment}By\index{tables!position on line/page}
writing\texttt{ }{\ttfamily\verb`valign=<char>`}\index{valign@\texttt{valign}}
where \verb`<char>` is one of \verb`t`, \verb`m` or \verb`b` the
vertical alignment of the table can be set relative to the text baseline:\texttt{ }\verb`t`\texttt{
}aligns the top of the table, \verb`b` the bottom of the table, and
\verb`m` the middle of the table with the text baseline. By default
\verb`valign=m` is set. Repeating an example from earlier (§\ref{sec:Row-variable-settings})
I have added letters A, B, C to show where the baseline is. Clearly
the top, middle and bottom of the respective tables aligns with the
baseline.
\begin{verbatim}
    A \tabulate[valign=t,rvar=x,rstep=0.2,rows=5]
      { \sin x/\cos x }[x=0][*] \quad
    B \tabulate[rspec={x,0.2,1/0.2}]
      { \tan x }[x=0][*] \quad
    C \tabulate[valign=b,rspec={x,0.2,5}]
      { \sqrt{\sec^2 x - 1} }[x=0][*]
\end{verbatim}
$\Longrightarrow$     A \tabulate[valign=t,rvar=x,rstep=0.2,rows=5,rstop=2]
      { \sin x/\cos x }[x=0.2][*] \quad
    B \tabulate[rspec={x,0.2,1/0.2}]
      { \tan x }[x=0.2][*] \quad
    C \tabulate[valign=b,rspec={x,0.2,5}]
      { \sqrt{\sec^2 x - 1} }[x=0.2][*]\medskip{}

As explained in §\ref{subsec:Adjoining-tables}, tables can be adjoined
to give the appearance of a single larger table. If tables with different
numbers of rows are adjoined in this manner, then a middle alignment
fails and a top alignment is necessary (so that the header rows of
the tables align).

\section{Formatting function values}

\label{sec:Function-value-formatting}\index{formatting!function values|(}\index{function-value formatting|(}In
previous tables in this document, function values have generally been
limited to a fairly narrow range of values. What happens when they
span orders of magnitude? Can scientific notation, expressly designed
to cope with such differing orders of magnitude, be accommodated in
a table in a natural way? Can rows or columns \textendash{} or individual
cells \textendash{} be rounded to different rounding values? Can we
indicate differences? Or form tables of function values in fraction
form? These and similar questions are our concern here.

\subsection{Trailing optional argument}

\label{subsec:Trailing-optional-argument}The primary tool for function-value
formatting is the trailing optional argument of the \verb`\tabulate`
command where the rounding value\index{rounding value} is specified,
padding with zeros is set or not (generally \emph{set} in tables),
scientific notation is set or not, and fraction-form output can be
specified. 

\subsubsection{Fraction-form output}

\begin{table}[t]
\centering
\centering{}\caption{\label{tab:Function-value-formatting}Formatting function values}
\begin{center}
\begin{tabular}{cc>{\raggedright}p{5cm}c}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small initial}\tabularnewline
\midrule
{\small\texttt{(pad)}} & {\small int} & {\small t-notation phantom padding} & \tabularnewline
{\small\texttt{signs}} & {\small int} & {\small sign handling for function values} & {\small\texttt{0}}\tabularnewline
{\small\texttt{diffs}} & {\small int} & {\small insert differences, pre-pad with $0$s} & {\small\texttt{0}}\tabularnewline
{\small\texttt{round}} & {\small tokens} & {\small row/col. dependent rounding value} & \tabularnewline
{\small\texttt{Q?}} & {\small tokens} & {\small special cell conditional} & \tabularnewline
{\small\texttt{A!}} & {\small tokens} & {\small special cell formatting} & \tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\end{table}
\label{subsec:Fraction-form-output}\index{fraction form!function values|(}\index{function-value formatting!fraction form|(}Function
values in a table can be presented in fraction form, but such output
requires far more computation than other forms since finding denominators
at the specified accuracy is an iterative process that needs doing
for every function value. But for small tables it is feasible. In
the tables below, approximations to small positive and inverse powers
of $\pi$ are listed to $2$ and $4$ decimal places of accuracy.
All the powers listed can be approximated to $4$-place accuracy by
$3$-figure denominators (and $\pi^{2}$ by a $2$-figure denominator).\index{row variable!verbatim values|(}\index{sfrac@\texttt{\textbackslash sfrac}}
\begin{verbatim}
    \def\mypi{\pi,\pi^2,\pi^3,\pi^{\sfrac12},\pi^{\sfrac13}}
    \tabulate[rdata=\mypi,rverb=1,rpos=1,rvar=k,ralign=l,
              chead={\small $2$ places}]{ k }[2/s] \qquad
    \tabulate[rdata=\mypi,rverb=1,rpos=1,rvar=k,ralign=l,
              chead={\small $4$ places}]{ k }[4/s]
\end{verbatim}
$\Longrightarrow$ \def\mydataiii{\pi,\pi^2,\pi^3,
    \pi^{\sfrac12},\pi^{\sfrac13}} 
    \tabulate[rdata=\mydataiii,rverb=1,rpos=1,
        rvar=k,ralign=l,chead={\small $2$ places}] 
     { k }[2/s]\qquad
\tabulate[rdata=\mydataiii,rverb=1,rpos=1,
        rvar=k,ralign=l,chead={\small $4$ places}] 
     { k }[4/s]\medskip{}

A second example shows that all four of \texttt{numerica}'s built-in
constants and their first few inverse powers can be approximated to
$5$ decimal places with $3$-figure denominators:\index{sfrac@\texttt{\textbackslash sfrac}}
\begin{verbatim}
    \tabulate[rdata={\pi,e,\phi,\gamma},rverb=1,/max=1000,
        rvar=k,cspec={n,1,4},chstyle=3,chnudge=9,rules=TthB,
        ctitle=\abs{k^{\sfrac1n}-p/q}<0.5\times10^{-5}] 
      { k^{\sfrac1n} }[n=1,k=1][/s5]
\end{verbatim}
$\Longrightarrow$     \tabulate[rdata={\pi ,e,\phi,\gamma},rverb=1,rvar=k,
        cspec={n,1,4},chstyle=3,chnudge=9,rules=TthB,/max=1000,
        ctitle=\lvert k^{\sfrac1n}-p/q \rvert<0.5\times10^{-5}] 
      { k^{\sfrac1n} }[n=1,k=1][/s5]\index{function-value formatting!fraction form|)}\index{fraction form!function values|)}\index{row variable!verbatim values|)}

\subsubsection{Scientific notation}

\index{function-value formatting!scientific notations|(}Scientific
notation\index{scientific notation} \textendash \index{scientific notation|seealso{t-notation}}
put \verb`x` in the trailing optional argument \textendash{} is generally
inappropriate for use in tables; see the first table below. Entering
\verb`xx` (second table) so that the notation extends to numbers
in the range $[1,10)$ helps, particularly with the \emph{left} alignment
of the function values, but the result is wasteful of space and the
repetition of the `$\times10$' is distracting and would be more
so for a larger table. The \verb`x` specification should be used
in tables, if at all, only for \emph{small} tables and special cases.
The \verb`t` option is much preferred; see §\ref{subsec:t-option}
following.
\begin{verbatim}
    \tabulate[rspec={x,1,6,0},chnudge=58]
      { e^x}[x=-2][*x]\qquad
    \tabulate[rspec={x,1,6,0},calign=l,chnudge=45]
      { e^x}[x=-2][*xx]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,1,6,0},chnudge=58]
      { e^x}[x=-2][*x]\qquad
    \tabulate[rspec={x,1,6,0},calign=l,chnudge=45]
      { e^x}[x=-2][*xx]

\subsection{The \texttt{t} option}

\label{subsec:t-option}\emph{HMF}\index{t-notation|(} uses a special
notation\index{function-value formatting|(seealso{t-notation}} for
coping with function values spanning different orders of magnitude.
This notation can be invoked by inserting \verb`t` in the trailing
optional argument. For the previous two tables the notation gives
a more compact and visually appealing result:
\begin{verbatim}
    \tabulate[rspec={x,1,2*3+1,0},chnudge=24]
      { e^x}[x=-3][*t]\qquad
    \tabulate[rspec={x,1,2*3+1,0},chnudge=24]
      { e^x}[x=-3][*tt]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,1,2*3+1,0},chnudge=24]
      { e^x}[x=-3][*t] \qquad
    \tabulate[rspec={x,1,2*3+1,0},chnudge=24]
      { e^x}[x=-3][*tt]\medskip{}

Doubling the \verb`t` in the trailing optional argument (second table)
extends the notation to numbers in the range $[1,10)$ \textendash{}
as doubling \verb`x` extends normal scientific notation to this range.

\subsubsection{Padding the exponent: \texttt{(pad)}}

In\index{padding with phantoms|(} the second table one might quibble
at the lack of alignment of the left parentheses. \emph{HMF} tends
to align these and \verb`numerica-tables` offers the setting 
\begin{verbatim}
    (pad) = <integer>
\end{verbatim}
to achieve the effect. (The parentheses are part of the key \textendash{}
a reminder of the \verb`t`-form of scientific notation.) \verb`<integer>`
is the number of digits/characters to pad to. Repeating the last two
tables with the setting \verb`(pad)=2` produces the following results:
\begin{verbatim}
    \tabulate[rspec={x,1,2*3+1,0},chnudge=24,(pad)=2]
      { e^x}[x=-3][*t]\qquad
    \tabulate[rspec={x,1,2*3+1,0},chnudge=24,(pad)=2]
      { e^x}[x=-3][*tt]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,1,2*3+1},rround=0,
                       chnudge=24,(pad)=2]
      { e^x}[x=-3][*t]\qquad
    \tabulate[rspec={x,1,2*3+1},rround=0,
                       chnudge=24,(pad)=2]
      { e^x}[x=-3][*tt]\medskip{}

\noindent Examples in \emph{HMF} of the style exemplified by the first
table are, among others, Tables 8.6, 9.2, 20.1, and of the style exemplified
by the second table, among many, Tables 9.9, 10.5, 13.1, 14.1, 19.1.\index{padding with phantoms|)}
But note:
\begin{itemize}
\item the \verb`(pad)` setting is relevant only when the \verb`t` option
is used in the trailing number-format argument of the \verb`\tabulate`
command and should have no effect otherwise.
\end{itemize}

\subsubsection{Accommodating signs in the \texttt{t}-notation}

\label{subsec:Signs}Instead\index{t-notation!accommodating signs}\index{function-value formatting!sign handling|(}
of $e^{x}$ as the test function, use $e^{x}-1.$ Now there are positive,
zero and negative function values to contend with. Recall that in
the \verb`t`-notation the \emph{exponent} is the parenthesized integer
part of a number and the \emph{significand} the following decimal
figures. \verb`numerica-tables` offers the \verb`signs`\index{signs@\texttt{signs}|(}
key to align (or not) the exponents. The setting is
\begin{verbatim}
    signs = <integer>
\end{verbatim}
Besides the do-nothing default (\verb`signs=0` or omitted), there
are four effective values for \verb`<integer>`:
\begin{itemize}
\item \verb`signs=2`\texttt{ }inserts a $+$ sign between exponent and
significand of every non-negative number;
\item \verb`signs=1`\texttt{ }inserts a $+$ sign between exponent and
significand of every non-negative number that immediately precedes
or follows a negative number;
\item \verb`signs=-1`\texttt{ }inserts a $+$ sign between exponent and
significand of any non-negative number that immediately precedes or
follows a negative number, and inserts a \emph{phantom} $+$ sign
between exponent and significand of every other non-negative number;
\item \verb`signs=-2`\texttt{ }inserts a \emph{phantom} $+$ sign between
exponent and significand of every non-negative number;
\end{itemize}
In the following examples, \verb`signs=-2`, \verb`signs=-1` and
\verb`signs=2`, all give acceptable results.
\begin{verbatim}
    \tabulate[rspec={x,1,2*3+1,0},(pad)=2,signs=-2]
      { e^x-1}[x=-3][4*tt] \qquad
    \tabulate[rspec={x,1,2*3+1,0},(pad)=2,signs=-1]
      { e^x-1}[x=-3][4*tt] \qquad
    \tabulate[rspec={x,1,2*3+1,0},(pad)=2,signs=2]
      { e^x-1}[x=-3][4*tt]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,1,2*3+1},rround=0,
        (pad)=2,signs=-2]
      { e^x-1}[x=-3][4*t']\qquad
    \tabulate[rspec={x,1,2*3+1},rround=0,
        (pad)=2,signs=-1]
      { e^x-1}[x=-3][4*t']\qquad
    \tabulate[rspec={x,1,2*3+1},rround=0,
        (pad)=2,signs=2]
      { e^x-1}[x=-3][4*t']\medskip{}

\noindent In \emph{HMF} Table 23.2 illustrates \verb`signs=-2`; Tables
10.1, 13.1, 14.1, 19.1 among many others illustrate \verb`signs=-1`;
and Tables 9.4, 10.6, 20.2, 22.11 among others illustrate \verb`signs=2`.

For \verb`signs=1`, see the second table in the next example.\index{t-notation|)}\index{function-value formatting!scientific notations|)}\index{padding with phantoms|)}\medskip{}


\subsection{Indicating signs outside the t-notation}

The \verb`signs` key is not limited to the \verb`t`-notation. In
the following tables where the notation is not used, positive values
for the key, including \verb`signs=1`\index{signs@\texttt{signs}},
give good results. The third table, corresponding to \verb`signs=0`,
shows that the comment extends to \emph{non-negative} values of the
\verb`signs` key.
\begin{verbatim}
    \tabulate[rspec={x,0.1,7},(pad)=2,signs=2]
      { 10\sin 5x}[x=-0.3][*4]\qquad
    \tabulate[rspec={x,0.1,7},(pad)=2,signs=1]
      { 10\sin 5x}[x=-0.3][*4]\qquad
    \tabulate[rspec={x,0.1,7},(pad)=2]
      { 10\sin 5x}[x=-0.3][*4]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.1,7},(pad)=2,signs=2]
      { 10\sin 5x}[x=-0.3][*4]\qquad
    \tabulate[rspec={x,0.1,7},(pad)=2,signs=1]
      { 10\sin 5x}[x=-0.3][*4]\qquad
    \tabulate[rspec={x,0.1,7},(pad)=2]
      { 10\sin 5x}[x=-0.3][*4]\medskip{}

\noindent\emph{HMF} seems to use \verb`signs=2` when the sign of
the function values changes every few entries and \verb`signs=1`
when there are runs of entries of the same sign. Over the range tabulated
here for $10\sin5x$, they would use the middle table of the three,
\verb`signs=1`.\index{function-value formatting!sign handling|)}\index{signs@\texttt{signs}|)}

\subsection{Cell-, row-, column-dependent rounding}

\label{subsec:Rounding-varying} \index{rounding value!row/col.dependent@row/col.\ dependent|(}In
§\ref{subsec:Fraction-form-output}, we created two tables of fraction-form
approximations to simple powers of $\pi$, one accurate to two places
of decimals, one to four. From version 3.1 (as distinct from version
3.0) \texttt{numerica-tables} offers the means of producing tables
with rounding values depending on position in the table. This is effected
through the key \verb`round`\index{round@\texttt{round}|(} which
sets the rounding value as a function of row and column variables.
In practice this usually means dependence on row or column variable
alone rather than both. In the example below, the rounding equals
the row variable value, \verb`round=r`, producing fractional approximations
to simple powers of $\pi$ at rounding values from $1$ to $5$, and
learn that all these powers can be approximated to $5$ decimal places
with $3$ figure denominators \textendash{} $\pi^{2}$ only just.
(The command \verb`\abs` used in \verb`ctitle` is defined in \texttt{numerica}.)\index{multi-function tables}\index{fraction form!row/col.-dependent rounding}
\begin{verbatim}
     \tabulate[ff,rspec={r,1,5},round=r,/max=999,chstyle=2,
            ctitle=\abs{\pi^k-\sfrac mn}<0.5\times10^{-r}]
        { \pi,\pi^2,\pi^3,\pi^{1/2},\pi^{1/3}}[r=1][/s]
\end{verbatim}
$\Longrightarrow$      \tabulate[ff,rspec={r,1,5},round=r,/max=999,chstyle=2,
              ctitle=\abs{\pi^k-\sfrac mn} < 0.5
                \times10^{-r}]
       { \pi ,\pi^2,\pi^3,\pi^{1/2},\pi^{1/3}}[r=1][/s]\index{fraction form!function values}\index{function-value formatting!fraction form}\medskip{}

Another place where a variable rounding value can be of value is when
a function being tabulated changes slowly for each step in the row
variable value. The value of the cosine for instance changes from
$1.0000$ to $\eval{\cos10\degree}[4]$ between $0\degree$ and $10\degree$.
\emph{Part} of a table of the cosine might be something like the following,
where values in the initial rows of the table are rounded to a higher
value than in later rows. In the example \verb`round` is set to an
expression in the row variable \verb`\theta` involving the boolean
expressions \verb`\theta<11` and \verb`\theta>10` which evaluate
to $0$ or $1$ depending on the value of \verb`\theta`. Thus \verb`round`
takes the value $6$ for the initial rows of the table and the value
$4$ thereafter. 
\begin{verbatim}
    \tabulate[o,rspec={\theta,1,6},calign=l,chnudge=15,
              round=6(\theta<11)+4(\theta>10)] 
      { \cos\theta }[\theta=8]
\end{verbatim}
$\Longrightarrow$     \tabulate[o,rspec={\theta,1,6},calign=l,chnudge=15,
              round=6(\theta<11)+4(\theta>10)] 
      { \cos\theta }[\theta=8]\index{rounding value!row/col.dependent@row/col.\ dependent|)}\index{round@\texttt{round}|)}

\subsection{Differences: \texttt{diffs}}

In\index{differences|(} fine-grained tables where function values
change only slowly from entry to entry it can be helpful to include
a difference entry between function-value entries as an aid to interpolation
(and a test of eyesight). By entering 
\begin{lyxcode}
diffs~=~<non-negative~integer>
\end{lyxcode}
the \verb`\tabulate` command will include differences in a table.
The \texttt{<non-negative integer>} is the maximum number of digits
in a difference. 
\begin{verbatim}
    \tabulate[rspec={x,0.01,6,2},diffs=3,
                rhnudge=9,chnudge=21]
        { \sinh x }[x=1,k=1][*4] \qquad
    \tabulate[rspec={x,0.01,6,2},diffs=2,
                rhnudge=9,chnudge=25]
        { \sinh x }[x=1][*4]\qquad
    \tabulate[rspec={x,0.01,6,2},diffs=4,
                rhnudge=9,chnudge=30]
        { \sinh x }[x=1][*4]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.01,6,2},diffs=3,
                rhnudge=9,chnudge=21]
        { \sinh x }[x=1,k=1][*4] \qquad
    \tabulate[rspec={x,0.01,6,2},diffs=2,
                rhnudge=9,chnudge=25]
        { \sinh x }[x=1][*4]\qquad
    \tabulate[rspec={x,0.01,6,2},diffs=4,
                rhnudge=9,chnudge=30]
        { \sinh x }[x=1][*4]\medskip{}

I have deliberately chosen the settings in the first table \textendash{}
\texttt{diffs=3} \textendash{} to give a good result. With the default
right alignment of the function-value columns, it is easy to get this
wrong. The evidence will be either in the misalignment of the first
row of function values (second table) or unnecessary padding of differences
with leading zeros (third table). It is a good idea to create your
table first, see how function values change between successive rows
and judge how many digits there will be in a difference. When the
\texttt{diffs} setting is too small, function values in the first
row are misaligned, the amount depending on how much too small. (A
left alignment of the function-value column is another way of tackling
this issue.) When the \texttt{diffs} setting is too big, alignment
is fine but differences are padded with unnecessary leading zeros,
meaning the column header will need a bigger nudge to bring \emph{it}
into alignment.

In the next example the function is \emph{decreasing}, showing how
it is the \emph{absolute value of the difference} between successive
function values that is tabulated. A difference is always a non-negative
value.
\begin{verbatim}
    \tabulate[rspec={x,0.01,6,2},diffs=2,
                rhnudge=9,chnudge=21]
      { e^{-x^2} }[x=0.05][*4]
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.01,6,2},diffs=2,
                rhnudge=9,chnudge=21]
      { e^{-x^2} }[x=0.05][*4] \index{differences|)}

\subsection{Formatting special values: \texttt{Q?} and \texttt{A!}}

\label{subsec:QA-formatting}You\index{function-value formatting!highlighting special values|(}\index{Q?@\texttt{Q?}|(}\index{A@\texttt{A"!}|(}
may wish to highlight or display in some special way a particular
function value or values. \verb`\nmcTabulate` has two related settings
that enable this: \verb`Q?=<tokens>` and \verb`A!=<tokens>`. As
the names suggest: Question? and Answer! 

The question should be an expression that \texttt{l3fp} can digest
and produce a boolean answer to (1 for `true', 0 for `false').
\emph{This is not a \LaTeX{} expression but an }\verb`l3fp` \emph{expression}.\footnote{Documentation about \texttt{l3fp }can be found in \texttt{interface3.pdf},
which is part of the \texttt{l3kernel} bundle.} For the user it should suffice to know that an expression formed
from decimal numbers (but \verb`l3fp` knows only the decimal dot),
parentheses \verb`( )`, the familiar arithmetic symbols, \verb`+`,\texttt{
}\verb`-`,\texttt{ }\verb`*`,\texttt{ }\verb`/` and \verb`^`,
relation symbols \verb`<`,\texttt{ }\verb`>`,\texttt{ }\verb`=`
and combinations like \verb`!=` (for $\ne$),\texttt{ }\verb`>=`
(for $\ge$), and\texttt{ }\verb`<=` (for $\le$) will be digested
by \verb`l3fp`. In addition there are \verb`||` for logical Or,
\verb`&&` for logical And, and \verb`!` for logical Not; \verb`exp(1)`
for $e$ and \verb`pi` (no backslash) for $\pi$. \texttt{numerica-tables}
provides \verb`MAX`\index{MAX@\texttt{MAX}} and \verb`MIN`\index{MIN@\texttt{MIN}}
for the maximum and minimum function values tabulated, and uses \verb`@`
to denote the function value of the current cell. 

So, a query might be \verb`Q?=@<0`, \emph{Is the current function
value negative?}, or \verb`Q?={@>=pi}`, \emph{Is the current function
value greater than or equal to $\pi$?} (The braces hide the equality
sign.) \verb`Q?={@=MIN}` (again note the braces) is the question:
\emph{Is the current function value equal to the minimum function
value for the whole table?}

The answer must be in the form of a \LaTeXe{} formatting statement,
again using \verb`@` to denote the function value of the current
cell. Thus \verb`A!=\mathbf{@}` is a valid answer; so is \verb`A!=\color{red}{@}`
(provided you have \verb`\usepackage{color}` in the preamble); and
so is \verb`A!=(@)`. Another valid answer is \verb`A!= ` , meaning
that function values satisfying the \verb`Q?` question are omitted
from the output. 

This can be useful to suppress `irrelevant' values in particular
contexts. For example, the \emph{non-zero} values of \verb`\binom{n}{m}`,
displaying as $\binom{n}{m}$, are the ones of interest. Rather than
cluttering the table with $0$s, suppress them, first by finding them
(\verb`Q?={@=0}`), and then by excising them (\verb`A!= ` ). The
table suppresses display of the header row with the \verb`headless`\index{headless@\texttt{headless}}
setting and instead displays it in the footer with the \verb`foot=''`\index{foot@\texttt{foot}}\index{footer!copies header}
setting, automatically changing what would have been $n\:\!\backslash\:\!m$
in the header to $n\:\!/\:\!m$ in the footer.
\begin{verbatim}
    \tabulate[rspec={n,1,5},rhnudge=-18,rfont=bf,chfont=bf,
              cspec={m,1,5},ctitle=\binom nm\qquad,
              headless,rules=f',foot='',Q?={@=0},A!=]
       { \binom{n}{m} }[n=1,m=1]
\end{verbatim}
$\Longrightarrow$      \tabulate[rspec={n,1,5},rhnudge=-18,rfont=bf,chfont=bf,
              cspec={m,1,5},ctitle=\binom nm\qquad,
              headless,rules=f',foot='',Q?={@=0},A!=]
       { \binom{n}{m} }[n=1,m=1]

\subsubsection{Star option: \texttt{\textbackslash nmcTabulate{*}}}

\label{sec:Star-option}If\index{nmcTabulate@\texttt{\textbackslash nmcTabulate}!star (*) option@star (\texttt{{*}}) option}
the \verb`Q?` question is satisfied by at least one function value
then adding a star (asterisk) to the \verb`\tabulate` command will
display the first such instance. Like other starred commands in the
\verb`numerica` suite (\verb`\eval*`, \verb`info*`, \verb`\macros*`,
\verb`\constants*`, \verb`\iter*`, \verb`\solve*` and \verb`\recur*`),
 \verb`\tabulate*` outputs a single number. Using the star means
you do not need an answering \verb`A!` to the query \verb`Q?` since
no formatting of table values is involved. 
\begin{verbatim}
    \tabulate*[rspec={n,1,12},cspec={m,1,4)},
               Q?={@<-1e-14||@>0.5+1e-14}]
      { \cos(m\pi/n) }[n=4,m=2][*4]
\end{verbatim}
$\Longrightarrow$      \tabulate*
      [rspec={n,1,12},cspec={m,1,4},
           Q?={@<-1e-14||@>0.5+1e-14}]
      { \cos(m\pi/n) }[n=4,m=2][*4]. Indeed, if you omit the \verb`Q?` setting from the previous table
so that all function values are visible then this is the value that
follows $0.5000$ in the \verb`m=2` column, the first function value
encountered outside the interval $[0,0.5]$. (The \verb`1e-14` is
to ensure rounding errors don't give a spurious result.)

If you want the \emph{maximum} value that has been tabulated then,
from version 3 of \texttt{numerica-tables}, you do not even need the
query: when \verb`\tabulate` is starred, \verb`Q?` is initialized
behind the scenes to \verb`@=MAX`\index{MAX@\texttt{MAX}}. Thus,
repeating the example from §\ref{sec:Column-variable-settings}, a
visual check of the table shows that the starred command has, indeed,
isolated the maximum value displayed:
\begin{verbatim}
    \tabulate[rspec={x,0.2,6},cspec={k,2,4},chnudge=27]
      { \sin kx }[k=3,x=0] $\longrightarrow
    \tabulate*[rspec={x,0.2,6},cspec={k,2,4},chnudge=27]
      { \sin kx }[k=3,x=0] $
\end{verbatim}
$\Longrightarrow$     \tabulate[rspec={x,0.2,6},cspec={k,2,4},chnudge=27]
      { \sin kx }[k=3,x=0.2][*] $\longrightarrow 
    \tabulate*[rspec={x,0.2,6},cspec={k,2,4},chnudge=27]
      { \sin kx }[k=3,x=0.2]$

\paragraph{Package option}

There\index{package options} is a package option \verb`Q?*`\index{Q?*@\texttt{Q?{*}}}
available which can be set to give some other default query than \verb`@=MAX`
if you so wish. For instance 
\begin{verbatim}
    \usepackage[Q?*={@=MIN}]{numerica-tables}
\end{verbatim}
makes `what is the minimum value tabulated?' the default query for
a \verb`\tabulate*` command. Or, \verb`Q?*=@>0` makes `what is
the first positive value encountered?' the default query for a \verb`\tabulate*`
command, and so on.

\paragraph{Errors}

If\index{error messages} \emph{no} function value satisfies a query
then a message is generated:
\begin{verbatim}
    \tabulate*[rspec={n,1,12)},cspec={m,1,4},Q?=@>1]
      { \cos(m\pi/n) }[n=4,m=2][*4]
\end{verbatim}
$\Longrightarrow$      \tabulate*
      [rspec={n,1,1+(15-4)},cspec={m,1,1+(5-2)},
           Q?=@>1]
      { \cos(m\pi/n) }[n=4,m=2][*4]

\noindent Note that the same message is presented if the error arises
from the package option query, which may well be confusing (for a
moment) since there is no explicit query present in the settings option.\index{Q?@\texttt{Q?}|)}\index{A@\texttt{A"!}|)}

\paragraph{Scientific notation}

If you want the number output in scientific notation when the star
option is chosen, then enter the exponent mark in the trailing number-format
option of the \verb`\tabulate*` command. This is straightforward
for a mark like the commonly used letter \verb`e`, but remember that
if you enter \verb`x` you will need to place the \verb`\tabulate*`
command between math delimiters, otherwise the \verb`\times` symbol
resulting from the \verb`x` option will generate a \LaTeX{} error
(`Missing \$ inserted'). Remembering the \verb`$` signs gives,
e.g.,
\begin{verbatim}
    $ \tabulate*[rspec={n,1,12},cspec={m,1,4},
                 Q?={@<-1e-14||@>0.5+1e-14}]
        { \cos(m\pi/n) }[n=4,m=2][*4x] $
\end{verbatim}
$\Longrightarrow$     $ 
    \tabulate*[rspec={n,1,12},cspec={m,1,4},
                 Q?={@<-1e-14||@>0.5+1e-14}]
      { \cos(m\pi/n) }[n=4,m=2][*4x]
    $.\index{formatting!function values|)}\index{function-value formatting|)}\index{function-value formatting!highlighting special values|)}

\section{Other matters}

Here I group items that do not fit naturally into the earlier categories.

\subsection{Nesting}

\label{subsec:Nesting}A\index{nmcTabulate@\texttt{\textbackslash nmcTabulate}!nesting}
\verb`\tabulate` command can be nested within other commands from
the \verb`numerica` suite, and those other commands can be nested
within a \verb`\tabulate` command.

Occasionally one might want to extract a value from a table to insert
in another command. This can be done by nesting a \verb`\tabulate*`
command with an appropriate \verb`Q?` setting within the other command.
In fact, from version 2 of \verb`numerica` on, the star is unnecessary.
All we require is that the \verb`Q?` setting is satisfied by at least
one tabulated function value. 
\begin{verbatim}
    \eval[env=$]{(\tabulate
      [rspec={n,1,15},cspec={m,1,5},Q?={@=MAX}]
        { \cos(m\pi/n) }[n=4,m=2][*4])\sinh t + 
          (\tabulate[rspec={n,1,15},cspec={m,1,5},Q?={@=MIN}]
             { \sin(m\pi/n) }[n=4,m=2][*4])\cosh t }
                [t=2][4]
\end{verbatim}
$\Longrightarrow$     \eval[env=$]{(\tabulate
      [rspec={n,1,15},cspec={m,1,5},Q?={@=MAX}]
        { \cos(m\pi/n) }[n=4,m=2][*4])\sinh t + 
          (\tabulate[rspec={n,1,15},cspec={m,1,5},Q?={@=MIN}]
             { \sin(m\pi/n) }[n=4,m=2][*4])\cosh t }
                [t=2][4]. 

Forming the table
\begin{verbatim}
     \tabulate[rspec={n,1,15,0},rpos=2,rules=Tth,
               cspec={m,1,5},ctitle=*,chstyle=2]
       { \cos(m\pi/n) }[n=4,m=2][*4]
\end{verbatim}
for the cosine (not presented here) and the table
\begin{verbatim}
    \tabulate[rspec={n,1,15,0},rpos=2,rules=Tth,
              cspec={m,1,5},chstyle=2,ctitle=*]
      { \sin(m\pi/n) }[n=4,m=2][*4]
\end{verbatim}
for the sine (also not presented here) and checking the entries shows
that indeed the maximum and minimum values are $0.9397$ and $-1.0000$
respectively.

If the \verb`Q?` setting is not satisfied by any function value a
familiar error message is shown \textendash{} with a tweak:
\begin{verbatim}
    \eval{$ (\tabulate
      [rspec={n,1,15},cspec={m,1,5},Q?=@>2]
         { \cos(m\pi/n) }[n=4,m=2][*4])\sinh t 
          $}[t=2][4]
\end{verbatim}
$\Longrightarrow$     \eval{$ 
      (\tabulate
        [rspec={n,1,15},cspec={m,1,5},
          Q?=@>2] 
            { \cos(m\pi/n) }[n=4,m=2][*4])\sinh t 
          $}[t=2][4]

\noindent Here, the \verb`(2)` tells us that the message refers to
a command at the second level, a \emph{nested} command.

A more likely situation is to want to nest other commands within a
\verb`\tabulate` command. I give an example in the documentation
to the associated package \verb`numerica-plus` about timing of signals
between points fixed on a rotating disk. 

\subsection{Saving tables to file}

\label{sec:reuse-setting}In\index{tables!saving to file}\index{saving tables to file}
earlier versions of \texttt{numerica-tables} it was possible to save
a table to file, or a row or a column or a particular value from a
table, by giving a \emph{setting} \verb`reuse` a value. From version
3.0.0, in the interests of simplifying use (and avoiding code complications)
the \verb`reuse` \emph{setting} has been discontinued. The \verb`\reuse`
(or \verb`\nmcReuse`) \emph{command} remains (as part of the \texttt{numerica}
package) and can be used to save the most recent table to file. 

In the following example, a table is created and then saved to file
and to the macro \verb`\mytable` by the subsequent \verb`\reuse`
command:
\begin{verbatim}
    \tabulate[rspec={x,0.25,5,2},rhnudge=9,rules=TthB,
              cspec={k,0.25,3,2},chstyle=2,ctitle=**]
      { a\sin kx }[a=2/\pi,{k}=3,{x}=0.25][*] 
    \reuse[renew]{mytable}
\end{verbatim}
$\Longrightarrow$     \tabulate
      [rspec={x,0.25,5},rround=2,
         rhead=x,ralign=r,rhnudge=9,
         cspec={k,0.25,3},chstyle=2,
         chround=2,calign=r,ctitle=**,rules=TthB]
      { a\sin kx }[a=2/\pi,{k}=3,{x}=0][*] 
    \reuse[renew]{mytable}\medskip{}

\noindent Now test the content of the control sequence \medskip{}

\noindent\verb`\mytable` $\Longrightarrow$ \mytable \medskip{}

\noindent Yes, \verb`\mytable` contains the table. The macro and
its contents have also been saved to file, the \texttt{.nmc} file
of the current document, hence \texttt{numerica-tables.nmc} in the
present instance. The contents of this file can be edited in a text
editor, or some limited file operations are possible with the \verb`\reuse`
command. These have been described in the associated document \texttt{numerica.pdf}.

\subsection{Viewing the \protect\LaTeX{} form}

\label{subsec:Viewing-latex-form}In\index{tables!viewing LaTeX form@viewing \LaTeX{} form}
previous versions of \texttt{numerica-tables} the \verb`dbg` and
\verb`view`\index{view@\texttt{view}} settings were disabled. In
version 3, they have been enabled to the extent that the \LaTeX{} form
of a table can be viewed by entering either \verb`dbg=11` or, less
nerdishly, \verb`view` into the settings option of \verb`\nmcTabulate`.
In the example a familiar table is specified and built in \LaTeX{}
and its form displayed with the \verb`view` setting:
\begin{verbatim}
    \tabulate[view,rspec={x,0.2,5}]
      { \sin x }[x=0.2]
\end{verbatim}
$\Longrightarrow$    \tabulate[view,rspec={x,0.2,5}]
      { \sin x }[x=0.2]

\chapter{Reference summary}

\section{Commands defined in \texttt{numerica-tables}}

\texttt{\textbackslash nmcTabulate, \textbackslash tabulate}

\subsubsection*{Package options}
\begin{itemize}
\item default rule configuration: \verb`rules`, \verb`norules`; see §§\ref{sec:Package-options},
\ref{subsec:Rules:-rules-setting}.
\item default query for \verb`\tabulate*`: \verb`Q?*`; see §\ref{sec:Star-option}.
\end{itemize}

\section{Settings for \texttt{\textbackslash nmcTabulate}}

\subsubsection*{Row variable specification: uniform case }

See §\ref{subsec:Row-var-spec-uniform}.
\begin{center}
\begin{center}
\begin{tabular}{l>{\raggedright}p{2cm}>{\raggedright}p{4cm}>{\raggedright}p{3cm}}
\toprule 
{\small key} & {\small type} & {\small meaning } & {\small comment}\tabularnewline
\midrule
{\small\texttt{rvar}} & {\small token(s)} & {\small row variable} & \tabularnewline
{\small\texttt{rstep}} & {\small real num.} & {\small step size} & \tabularnewline
{\small\texttt{rstop}} & {\small real num.} & {\small stop value} & {\small excludes }{\small\texttt{rows}}\tabularnewline
{\small\texttt{rows}} & {\small int} & {\small number of rows} & {\small excludes }{\small\texttt{rstop}}\tabularnewline
{\small\texttt{rspec}} & {\small comma list} & {\small\texttt{\{rvar}}{\small , }{\small\texttt{rstep}}{\small , }{\small\texttt{rows\}}} & {\small short form spec.}\tabularnewline
{\small\texttt{rround}} & {\small int} & {\small rounding} & {\small default: }{\small\texttt{1}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}\pagebreak{}
\par\end{center}

\subsubsection*{Row variable specification: non-uniform case}

See §\ref{subsec:Row-var-spec-nonuniform}.

\begin{center}
\begin{tabular}{l>{\raggedright}p{2cm}>{\raggedright}p{4cm}>{\raggedright}p{3cm}}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small comment}\tabularnewline
\midrule
{\small\texttt{rfunc}} & {\small token(s)} & {\small formula for row var. values} & \tabularnewline
{\small\texttt{rdata}} & {\small comma list or macro} & {\small list or macro (containing list) of row var. values} & \tabularnewline
{\small\texttt{rfile}} & {\small chars} & {\small filepath/filename} & {\small file contains list of row var. values}\tabularnewline
{\small\texttt{rdelim}} & {\small char} & {\small item separator for }{\small\texttt{rdata}}{\small{} or }{\small\texttt{rfile}}{\small{}
lists} & {\small defaults to }{\small\texttt{,}}{\small{} or }{\small\texttt{;}}{\small{}
depending as }{\small\texttt{.}}{\small{} or }{\small\texttt{,}}{\small{}
is decimal mark}\tabularnewline
{\small\texttt{rverb}} & {\small fp (}{\small\texttt{0/0.5/1}}{\small )} & {\small display }{\small\texttt{rdata}}{\small , }{\small\texttt{rfile}}{\small{}
values verbatim (}{\small\texttt{1}}{\small ), or slash fractions formatted
(}{\small\texttt{0.5}}{\small )} & {\small initialized to }{\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}

\subsubsection*{Row variable column formatting}

See §\ref{subsec:Row-var-col-formatting}.
\begin{center}
\begin{center}
\begin{tabular}{l>{\raggedright}p{2cm}>{\raggedright}p{4cm}>{\raggedright}p{2cm}}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small initial}\tabularnewline
\midrule
{\small\texttt{rpos}} & {\small int (}{\small\texttt{0}}{\small\ldots}{\small\texttt{4}}{\small )} & {\small column placement} & {\small\texttt{1}}\tabularnewline
{\small\texttt{ralign}} & {\small char (}{\small\texttt{r/c/l}}{\small )} & {\small horizontal alignment} & {\small\texttt{r}}\tabularnewline
{\small\texttt{rfont}} & {\small chars} & {\small font (}{\small\verb`\math<chars>`}{\small )} & \tabularnewline
{\small\texttt{rhead}} & {\small tokens} & {\small header} & \tabularnewline
{\small\texttt{rhnudge}} & {\small fp} & {\small nudge header }{\small{\small\verb`<fp>`}}{\small{} mu} & {\small\texttt{0}}\tabularnewline
{\small\texttt{rhsize}} & {\small int (}{\small\texttt{-4}}{\small\ldots}{\small\texttt{5}}{\small )} & {\small font size relative to }{\small\verb`0=\normalsize`} & {\small\texttt{0}}\tabularnewline
{\small\texttt{rmath}} & {\small char (}{\small\texttt{s/t/d}}{\small )} & {\small script-, text-, displaystyle} & {\small\texttt{t}}\tabularnewline
{\small\texttt{rvar'}} & {\small tokens} & {\small 2nd row variable col. spec.} & \tabularnewline
{\small\texttt{rhead'}} & {\small tokens} & {\small header of 2nd row var. col. (if it exists)} & \tabularnewline
{\small\texttt{rhnudge'}} & {\small fp} & {\small nudge 2nd row var. col. header }{\small{\small\verb`<fp>`}}{\small{}
mu} & {\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}\pagebreak{}
\par\end{center}

\subsubsection*{Column-variable specification }

See §\ref{sec:Column-variable-settings}.
\begin{center}
\begin{center}
\begin{tabular}{ll>{\raggedright}p{3cm}l}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small comment}\tabularnewline
\midrule
{\small\texttt{cvar}} & {\small token(s)} & {\small column variable} & \tabularnewline
{\small\texttt{cstep}} & {\small real num.} & {\small step size} & \tabularnewline
{\small\texttt{cstop}} & {\small real num.} & {\small stop value} & {\small either }{\small\texttt{cstop}}\tabularnewline
{\small\texttt{cols}} & {\small int} & {\small number of columns} & {\small or }{\small\texttt{cols}}\tabularnewline
{\small\texttt{cspec}} & {\small comma list} & {\small\texttt{\{cvar,cstep}}{\small ,}{\small\texttt{cols\}}} & {\small short form spec.}\tabularnewline
{\small\texttt{chround}} & {\small int} & {\small col. var. rounding} & {\small default: }{\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\par\end{center}

\subsubsection*{Column-variable header formatting}

See §\ref{subsec:Column-header-formatting}.
\begin{center}
\begin{center}
{\small{}%
\begin{tabular}{ll>{\raggedright}p{4cm}l}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small default}\tabularnewline
\midrule
{\small\texttt{chstyle}} & {\small int (}{\small\texttt{0}}{\small\ldots}{\small\texttt{4}}{\small )} & {\small header style} & {\small\texttt{0}}\tabularnewline
{\small\texttt{chead}} & {\small token(s)} & {\small user-defined col. var. header} & \tabularnewline
{\small\texttt{calign}} & {\small char (}{\small\texttt{r}}{\small /}{\small\texttt{c}}{\small /}{\small\texttt{l}}{\small )} & {\small column alignment} & {\small\texttt{r}}\tabularnewline
{\small\texttt{chnudge}} & {\small fp} & {\small nudge header }{\small{\small\verb`chnudge`}}{\small{}
mu} & {\small\texttt{0}}\tabularnewline
{\small\texttt{chfont}} & {\small chars} & {\small font (}{\small\verb`\math<chars>`}{\small )} & \tabularnewline
{\small\texttt{chmath}} & {\small char (}{\small\texttt{s}}{\small /}{\small\texttt{t}}{\small /}{\small\texttt{d}}{\small )} & {\small script-, text- or displaystyle} & {\small\texttt{t}}\tabularnewline
{\small\texttt{chsize}} & {\small int (}{\small\texttt{-4}}{\small\ldots}{\small\texttt{5}}{\small )} & {\small font size relative to }{\small\verb`0 => \normalsize`} & {\small\texttt{0}}\tabularnewline
\bottomrule
\end{tabular}}
\par\end{center}
\par\end{center}

\subsubsection*{Function-value formatting}

See §\ref{sec:Function-value-formatting}.
\begin{center}
\begin{center}
\begin{tabular}{cc>{\raggedright}p{5cm}c}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small initial}\tabularnewline
\midrule
{\small\texttt{(pad)}} & {\small int} & {\small t-notation phantom padding} & \tabularnewline
{\small\texttt{signs}} & {\small int} & {\small sign handling for function values} & {\small\texttt{0}}\tabularnewline
{\small\texttt{diffs}} & {\small int} & {\small insert differences, pre-pad with $0$s} & {\small\texttt{0}}\tabularnewline
{\small\texttt{round}} & {\small tokens} & {\small row/col. dependent rounding value} & \tabularnewline
{\small\texttt{Q?}} & {\small tokens} & {\small special cell conditional} & \tabularnewline
{\small\texttt{A!}} & {\small tokens} & {\small special cell formatting} & \tabularnewline
\bottomrule
\end{tabular}
\par\end{center}\pagebreak{}
\par\end{center}

\subsubsection*{Whole-of-table formatting }

See §\ref{sec:Whole-of-table-formatting}.
\begin{center}
\begin{center}
\begin{tabular}{ll>{\raggedright}p{4cm}l}
\toprule 
{\small key} & {\small type} & {\small meaning} & {\small initial}\tabularnewline
\midrule
{\small\texttt{ctitle}} & {\small token(s)} & {\small collective title for columns} & \tabularnewline
{\small\texttt{csubttl}} & {\small token(s)} & {\small subtitle row for} {\small function- value cols} & \tabularnewline
{\small\texttt{calign}} & {\small char(}{\small\texttt{r}}{\small /}{\small\texttt{c}}{\small /}{\small\texttt{l}}{\small )} & {\small column alignment} & {\small\texttt{r}}\tabularnewline
{\small\texttt{headless}} &  & {\small suppress header row} & \tabularnewline
{\small\texttt{foot}} & {\small token(s)} & {\small table-wide footer row} & \tabularnewline
{\small\texttt{rules}} & {\small char(s)} & {\small horizontal rule spec.} & {\small\texttt{ThB}}\tabularnewline
{\small\texttt{norules}} &  & {\small cancel all rules} & \tabularnewline
{\small\texttt{rpos}} & {\small int (}{\small\texttt{0}}{\small\ldots}{\small\texttt{4}}{\small )} & {\small row variable col. position(s)} & {\small\texttt{1}}\tabularnewline
{\small\texttt{rbloc}} & {\small integer comma list} & {\small row block specification} & \tabularnewline
{\small\texttt{rblocsep}} & {\small length} & {\small extra space between row blocks} & {\small\texttt{1ex}}\tabularnewline
{\small\texttt{transpose}} &  & {\small show funct. vals in rows} & \tabularnewline
{\small\texttt{valign}} & {\small char (}{\small\texttt{t}}/{\small\texttt{m}}/{\small\texttt{b}}{\small )} & {\small vertical alignment of table relative to text baseline} & {\small\texttt{m}}\tabularnewline
\bottomrule
\end{tabular}
\par\end{center}
\par\end{center}

\subsubsection*{Miscellaneous settings}
\begin{itemize}
\item \verb`view`, equivalent to \verb`dbg=11`: show the \LaTeX{} expression
for the table; see §\ref{subsec:Viewing-latex-form}.
\end{itemize}
{\small\printindex}{\small\par}
\end{document}