I am trying to produce a diagram similar to the following (from Wikipedia)


in TiKz. Now, I thought this would be relatively simple, however, I am having trouble getting the elements (the dots) to be of the correct size and am unsure how to elegantly surround the elements in an ellipse. My current TiKz code is


 \begin{tikzpicture}[ele/.style={fill=black,minimum size=2pt,circle}, node distance=7pt]
  \node[ele] (a1) {};
  \node[ele] (a2) [below=of a1] {};
  \node[ele] (a3) [below=of a2] {};
  \node[ele] (a4) [below=of a3] {};
  \node[ele] (b1) [right=of a1,xshift=15pt] {};
  \node[ele] (b2) [below=of b1] {};
  \node[ele] (b3) [below=of b2] {};
  \node[ele] (b4) [below=of b3] {};
  \draw[->,thick] (a1) -- (b4);
  \draw[->,thick] (a2) -- (b2);
  \draw[->,thick] (a3) -- (b1);
  \draw[->,thick] (a4) -- (b3);

which renders as


Can anyone recommend a more elegant means of drawing such a diagram?

  • 1
    Set inner sep in the ele style to 0.
    – Caramdir
    Jun 5, 2011 at 20:22

2 Answers 2


 \begin{tikzpicture}[ele/.style={fill=black,circle,minimum width=.8pt,inner sep=1pt},every fit/.style={ellipse,draw,inner sep=-2pt}]
  \node[ele,label=left:$a$] (a1) at (0,4) {};    
  \node[ele,label=left:$b$] (a2) at (0,3) {};    
  \node[ele,label=left:$c$] (a3) at (0,2) {};
  \node[ele,label=left:$d$] (a4) at (0,1) {};

  \node[ele,,label=right:$1$] (b1) at (4,4) {};
  \node[ele,,label=right:$2$] (b2) at (4,3) {};
  \node[ele,,label=right:$3$] (b3) at (4,2) {};
  \node[ele,,label=right:$4$] (b4) at (4,1) {};

  \node[draw,fit= (a1) (a2) (a3) (a4),minimum width=2cm] {} ;
  \node[draw,fit= (b1) (b2) (b3) (b4),minimum width=2cm] {} ;  
  \draw[->,thick,shorten <=2pt,shorten >=2pt] (a1) -- (b4);
  \draw[->,thick,shorten <=2pt,shorten >=2] (a2) -- (b2);
  \draw[->,thick,shorten <=2pt,shorten >=2] (a3) -- (b1);
  \draw[->,thick,shorten <=2pt,shorten >=2] (a4) -- (b3);

enter image description here

  • minimum width=.8pt,inner sep=1pt to get smaller points Jun 5, 2011 at 21:32
  • Any chance explaining this wall of code? Aug 9, 2017 at 11:34
  • Why did you hard code the coordinates, for example, (4,4), etc? The OP did not use such hardcoded coordinates. And I think using hardcoded coordinates makes future update much harder. What do you think? I found another related answer but without hardcoded coordinates here. Mar 31, 2021 at 11:33

Wallner's answer was one of the best mapping diagram I have ever seen. I modified his code for your problem

    % draw the sets
    \filldraw[fill=blue!20, draw=blue!60] (-1.5,0) circle (1cm);
    \filldraw[fill=red!20, draw=red!60] (1.5,0) circle (1cm);

    % the texts
    \node at (-1.5,1.5) {$X$};
    \node at (1.5,1.5) {$Y$};

    % the points in the sets (here I just create nodes to use them later on to position
    % the circles and the arrows
    \node (x1) at (-1.5,0.7) {$a$};
    \node (x2) at (-1.5,0.3) {$b$};
    \node (x3) at (-1.5,-0.2) {$c$};
    \node (x4) at (-1.5,-0.7) {$d$};
    \node (y1) at (1.5,0.7) {$1$};
    \node (y2) at (1.5,0.3) {$2$};
    \node (y3) at (1.5,-0.2) {$3$};
    \node (y4) at (1.5,-0.7) {$4$};

    % draw the arrows
    \draw[->] (x1) -- (y4);
    \draw[->] (x2) -- (y2);
    \draw[->] (x3) -- (y1);
    \draw[->] (x4) -- (y3);

\caption{Mapping diagram of relation $S$}

which produce the mapping below: enter image description here

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .