Problem declaring newcommand with tikzpicture inside

Question

I plan to put some common commutative diagrams and short exact sequences as new commands in my *.sty file. However, putting this code in my *.sty file

\ProvidesPackage{myLayout}
\usepackage{ifthen,amsopn}
\newcommand{\SES}[3]{
\begin{center}
\begin{tikzpicture}[every node/.style={midway}]
\matrix[column sep={2em},
        row sep={0em}] at (0,0)
{ \node(A) {$0$} ;
& \node(B) {#1} ;
& \node(C) {#2} ;
& \node(D) {#3} ;
& \node(E) {$0$} ; \\};
\draw[->] (A) -- (B)  node[anchor=south] {};
\draw[->] (B) -- (C)  node[anchor=south] {};
\draw[->] (C) -- (D)  node[anchor=south] {};
\draw[->] (D) -- (E)  node[anchor=south] {};
\end{tikzpicture}
\end{center}
}

Copy-pasting this whole thing to my *.tex document (except the \newcommand part) and replacing #1, #2, #3 with $A$, $B$ and $C$ gives me a wonderful exact sequence diagram.

I want to know what's going wrong when I paste the same exact code in \newcommand. Moreover, I'd like to know if there is a more efficient way to "save layouts" of tikzpicture so that I don't have to copy paste the same codes every time (especially if I'm just making minor adjustments).

The ultimate goal would be to just write one line like this

\SES{A}{B}{C}

for a short exact sequence.

A general and efficient solution would be a very nice plus.

Any help would be very much appreciated.

Answer 1

You can use ampersand replacement; however, I'd not use the center environment, for better flexibility (and use \[...\] instead, around the command). You can also simplify the code using matrix of math nodes.

\documentclass{article}
\usepackage{tikz}
\usetikzlibrary{matrix}

\newcommand{\SES}[3]{%
  \begin{tikzpicture}[every node/.style={midway},ampersand replacement=\&]
  \matrix [matrix of math nodes,column sep=2em, row sep=0em] at (0,0) {
  \node(A) {0} ;
  \& \node(B) {#1} ;
  \& \node(C) {#2} ;
  \& \node(D) {#3} ;
  \& \node(E) {0} ; \\};
  \draw[->] (A) -- (B)  node[anchor=south] {};
  \draw[->] (B) -- (C)  node[anchor=south] {};
  \draw[->] (C) -- (D)  node[anchor=south] {};
  \draw[->] (D) -- (E)  node[anchor=south] {};
  \end{tikzpicture}%
}

\begin{document}

\[
\SES{A}{B}{C}
\]

\end{document}

However, I'd recommend using tikz-cd.

\documentclass{article}
\usepackage{tikz-cd}

\newcommand{\SES}[3]{%
  \begin{tikzcd}[ampersand replacement=\&]
  0 \arrow[r] \& #1 \arrow[r] \& #2 \arrow[r] \& #3 \arrow[r] \& 0
  \end{tikzcd}%
}

\begin{document}

\[
\SES{A}{B}{C}
\]

\end{document}

With the simplified syntax of tikz-cd you'll probably give up to defining several types of commutative diagrams with macros.

Answer 2

The problem is that latex is converting the & before tikz can parse it. See example on page 1048.

\documentclass{article}
\usepackage{tikz}

\newcommand{\SES}[3]{%
\begin{center}
\begin{tikzpicture}[every node/.style={midway}]
\let\&=\pgfmatrixnextcell
\matrix[column sep={2em},
        row sep={0em}] at (0,0)
{ \node(A) {$0$} ;
\& \node(B) {#1} ;
\& \node(C) {#2} ;
\& \node(D) {#3} ;
\& \node(E) {$0$} ; \\};
\draw[->] (A) -- (B)  node[anchor=south] {};
\draw[->] (B) -- (C)  node[anchor=south] {};
\draw[->] (C) -- (D)  node[anchor=south] {};
\draw[->] (D) -- (E)  node[anchor=south] {};
\end{tikzpicture}
\end{center}
}

\begin{document}

\SES{$A$}{$B$}{$C$}

\end{document}