# Doing iterative arithmetic inside of an iterative tabularx

I'm trying to create a matrix table that will iteratively fill in the cells based on inputs that I can change. I originally tried the solution shown here, but I wasn't able to get Latex to properly expand when I was actually using the iterative value \i in the cells, rather than pure text like t1 and t2.

So I also came across this great solution here, which did work with inputting \i as a value. However, I'm having problems when doing mathematical operations on that \i and further values. In accordance with that second solution, here's my MWE:

\documentclass[]{article}
\usepackage{tikz}
\usepackage{tabularx}
\begin{document}
\newcommand\costsmall{60}
\newcommand\costmed{80}
\newcommand\costlarge{100}
\newcommand\rentsmall{10}
\newcommand\rentmed{15}
\newcommand\rentlarge{20}

\makeatletter
\newtoks\@tabtoks
\newcommand*\resettabtoks{\global\@tabtoks{}}
\newcommand*\printtabtoks{\the\@tabtoks}
\makeatother

\resettabtoks
\addtabtoks{\textbf{Num1} & \textbf{Num2} & \textbf{Num3} & \textbf{Num2/Num3}\\ \cline{1-4}}

\foreach \x [count=\xi] in {\costsmall,\costmed,\costlarge}{
\foreach \y [count=\yi] in {\the\numexpr\x - 10,\the\numexpr\x,\the\numexpr\x + 10}{
\foreach \z [count=\zi] in {\rentsmall,\rentmed,\rentlarge}{
\ifnum\zi=2 {
\ifnum\yi=2 {
}
\else {
}
\fi
}
\else {
}
\fi
%           \newcommand{\testthing}[2]{y is #1, z is #2, and together they are #1/#2}
%           \testthing{\y}{\z}

\ifnum\zi=3 {
\ifnum\yi=3 {
}
\else{
}
\fi
}
\fi
}
}
}

\begin{tabularx}{\textwidth}{|c|c|c|c|c|}
\printtabtoks
\end{tabularx}%

\end{document}


As a result, I get this plot, where the "# of rents to pay off" has clearly been calculated incorrectly for most of the cases:

In this code, this is at the \eaddtabtoks{\y/\z} line. I've been able to trace it back and find that it's failing in all the cases where \y has 10 subtracted or added to it in the \foreach line. In those cases, it looks like a strange order of operations takes hold, or the 10 is not seen as a number or something. I've played with the \y/\z line endlessly, trying different options of \the\numexpr and brackets and parentheses to get the order of operations done correctly, but nothing seems to work.

As a side note, I did have to use this janky \ifnum\zi=2 method of centering the first few multicolumns because I couldn't get multicolumns to work with this method. \multicolumn{9}{*}{random} worked fine with a string like random, but once I put in an iterator like \x or \y, I got errors that I couldn't fix.

• This is a general advice: if you provide a complete example that starts with \documentclass, ends with \end{document} and can be compiled, many more users will look at your code. – user121799 Jun 22 at 16:59
• tabularx can do nothing useful with a preamble of \begin{tabularx}{\textwidth}{|c|c|c|c|c|} just use tabular. The tabularx version is evaluated multiple times with trial column widths which makes embedding calculations interesting,m but you do not want tabularx here anyway. – David Carlisle Jun 22 at 17:12
• you are missing a space after 10 in that calculation – David Carlisle Jun 22 at 17:20
• Note that you should not enclose the “true text” and the “false text” in braces. – egreg Jun 22 at 20:33
• Can you explain what the idea is? How should the entries be computed? – egreg Jun 22 at 21:08