[Tex/LaTex] pgfplots: “group” legend entries

groupinglegendpgfplots

Actually, I'm working on a study where I've to plot a lot of temperature data etc. Therefore, I use "pgfplots". Everything works pretty well, so that I get the result as shown in the attached picture

Nodal Temperatures .

by using the following code:

\documentclass[tikz,border=1pt]{standalone}     
\usepackage[ngerman]{babel}                         
\usepackage[ansinew]{inputenc}
\usepackage[T1]{fontenc}                            
\usepackage{graphicx} 
\usepackage{pgfplots}
\usepackage{amsmath} 

% Settings for tikz-Image
\pgfplotsset{compat=1.10, 
   01_temp/.style={
      %title=Knotentemperaturen,
      legend columns=4,
      legend style={ 
         font=\scriptsize,
         legend pos=north east,
         draw=none,
         /tikz/column 2/.style={
            column sep=5pt,
         }
      }, 
      xmin=0,
      xmax=200,
      xlabel={Zeit $t$ [h]},
      x tick style={
        color=black,
        thin
      },      
      ymin=15,
      ymax=50,
      ylabel={Knotentemperatur $T$ [$^\circ$C]},
      y tick style={
        color=black,
        thin
      },
      height=9cm,
      width=15cm,
      grid=major,
      grid style={
         solid,
         ultra thin,
         gray
      },
      /pgf/number format/.cd,
      use comma,
      set thousands separator={},
   } 
} 

\begin{document} 

 \begin{tikzpicture} 
    \begin{axis}[01_temp]
       \addplot[black, thin, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70101} 
       \addplot[black, very thick, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70101} 
       \addplot[red, thin, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70201} 
       \addplot[red, very thick, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70201} 
       \addplot[green, thin, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70301}
       \addplot[green, very thick, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70301}
       \addplot[blue, thin, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70401}
       \addplot[blue, very thick, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70401}
       \addplot[pink, thin, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70501}
       \addplot[pink, very thick, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70501}
       \addplot[gray, thin, solid] table [x index=0,y index=1, skip first n=4] {01a_temp.dat};
       \addlegendentry{Klima}
       \addplot[gray, very thick, solid] table [x index=0,y index=1, skip first n=4] {01a_temp.dat};
       \addlegendentry{Klima}
    \end{axis} 
 \end{tikzpicture}

\end{document}

Both files (used in the code-example above) that contain the data look like (extract!):

      My office.
  My project.
  Nodal Temperature

  Table of function values
          t          K1          K2          K3          K4          K5
         [h]        [°C]        [°C]        [°C]        [°C]        [°C]
       0.000       20.00       20.00       20.00       20.00       20.00
       1.000       19.98       19.98       19.98       19.99       20.00
       2.000       19.95       19.95       19.95       19.98       19.99
       3.000       19.92       19.91       19.92       19.98       19.99
       4.000       19.94       19.94       19.94       20.04       20.04
       5.000       20.07       20.07       20.08       20.19       20.17
       6.000       20.33       20.33       20.33       20.46       20.41
       7.000       20.70       20.70       20.70       20.81       20.72
       8.000       21.16       21.16       21.17       21.22       21.08
       9.000       21.69       21.69       21.70       21.66       21.47
      10.000       22.27       22.27       22.28       22.12       21.88
      11.000       22.89       22.89       22.89       22.58       22.28
      12.000       23.52       23.52       23.51       23.01       22.66
      13.000       24.16       24.16       24.14       23.42       23.02
      14.000       24.80       24.79       24.76       23.80       23.35
      15.000       26.61       26.61       26.55       25.22       24.61
      16.000       28.55       28.55       28.44       26.64       25.85
      17.000       30.53       30.52       30.36       27.99       27.03
      18.000       31.27       31.26       31.06       28.14       27.14
      19.000       31.97       31.96       31.72       28.25       27.24
      20.000       32.63       32.62       32.32       28.33       27.30
      21.000       33.26       33.25       32.89       28.38       27.34
      22.000       33.86       33.84       33.43       28.41       27.36
      23.000       34.43       34.41       33.93       28.41       27.36
      24.000       34.98       34.95       34.40       28.40       27.34
      25.000       35.49       35.46       34.84       28.38       27.32
      26.000       35.98       35.95       35.26       28.34       27.28
      27.000       36.45       36.41       35.64       28.29       27.23
      28.000       36.89       36.84       36.00       28.23       27.17
      29.000       37.31       37.26       36.33       28.16       27.11
      30.000       37.71       37.65       36.64       28.09       27.04

Now to my question: If you look at the data-plot, you can see that I get always two lines (thin and thick) for every node. For example, Node "K 70101" has a thin and a thick data plot. In the example, both lines lie upon each other. (That's why it is a little bit difficult, to find the thin plots …).

Anyway: is it possible to generate a kind of "grouping" in the legend in that way that both lines (thin and thick) just get one legendentry? It's a little bit difficult to describe – that's why I created an image to show what I mean:

grouped legend

It would be nice, if someone could help me. Unfortunately I didn't found some examples that I can modify to get the result I need …

Thank you in advance!

So, during the night I got a little idea and made a small step closer to the result I want to get. I managed it to "group" the legend-entries – unfortunately in one single column:

one single column

The "new" code is the following:

\documentclass[tikz,border=1pt]{standalone}         
\usepackage[ngerman]{babel}                         
\usepackage[ansinew]{inputenc}                      
\usepackage[T1]{fontenc}                            
\usepackage{graphicx} 
\usepackage{pgfplots}
\usepackage{amsmath} 

\pgfplotsset{compat=1.10, 
   01_temp/.style={         
%%   legend columns=2,
     legend style={ 
       font=\scriptsize,
           legend pos=north east,
       draw=none,
%%           /tikz/column 2/.style={
%%            column sep=5pt,
%%         }
      }, 
      xmin=0,
      xmax=200,
      xlabel={Zeit $t$ [h]},
      x tick style={
        color=black,
        thin
      },      
      ymin=15,
      ymax=50,
      ylabel={Knotentemperatur $T$ [$^\circ$C]},
      y tick style={
        color=black,
        thin
      },
      height=9cm,
      width=15cm,
      grid=major,
      grid style={
         solid,
         ultra thin,
         gray
      },
      /pgf/number format/.cd,
      use comma,
      set thousands separator={},
   } 
} 

\begin{document} 
 \begin{tikzpicture} 
    \begin{axis}[01_temp]
       \addplot[black, thin, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \addlegendentry{} 
       \addplot[black, very thick, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \addlegendentry{\smash{\raisebox{-0.3ex}{K 70101}}} 
       \addplot[red, thin, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \addlegendentry{} 
       \addplot[red, very thick, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \addlegendentry{\smash{\raisebox{-0.3ex}{K 70201}}} 
       \addplot[green, thin, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \addlegendentry{}
       \addplot[green, very thick, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \addlegendentry{\smash{\raisebox{-0.3ex}{K 70301}}}
       \addplot[blue, thin, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \addlegendentry{}
       \addplot[blue, very thick, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \addlegendentry{\smash{\raisebox{-0.3ex}{K 70401}}}
       \addplot[pink, thin, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \addlegendentry{}
       \addplot[pink, very thick, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \addlegendentry{\smash{\raisebox{-0.3ex}{K 70501}}}
       \addplot[gray, thin, solid] table [x index=0,y index=1, skip first n=4] {01a_temp.dat};
       \addlegendentry{}
       \addplot[gray, very thick, solid] table [x index=0,y index=1, skip first n=4] {01a_temp.dat};
       \addlegendentry{\smash{\raisebox{-0.3ex}{Klima}}}
    \end{axis} 
 \end{tikzpicture}    
\end{document}

It would be fine if someone could help me, solving the remaining problems:

How is it now possible, to use a kind of "3 column"-layout with those grouped legend-entries?
How is it possible to reduce the vertical space of the 'sample-lines' in the legend?

So, I managed to move another step. My legend looks now like that:

enter image description here

… the code is the following:

    \documentclass[tikz,border=1pt]{standalone}     
\usepackage[ngerman]{babel}                         
\usepackage[ansinew]{inputenc}                  
\usepackage[T1]{fontenc}                            
\usepackage{graphicx} 
\usepackage{pgfplots}
\usepackage{amsmath} 

\pgfplotsset{compat=1.10, 
   01_temp/.style={
      title={\textbf{Knotentemperaturen, $\varphi(t) = 0$}},
      legend columns=6,
     legend style={ 
       font=\scriptsize,
           legend pos=north east,
       draw=none,
       /tikz/column 1/.style={
         yshift=-0.2em,
       },
       /tikz/column 2/.style={
         column sep=-2em,
       },
       /tikz/column 5/.style={
         column sep=-2em,
       },
       /tikz/column 9/.style={
         column sep=-2em,
       }
      }, 
      xmin=0,
      xmax=200,
      xlabel={Zeit $t$ [h]},
      x tick style={
        color=black,
        thin
      },      
      ymin=15,
      ymax=50,
      ylabel={Knotentemperatur $T$ [$^\circ$C]},
      y tick style={
        color=black,
        thin
      },
      height=9cm,
      width=15cm,
      grid=major,
      grid style={
         solid,
         ultra thin,
         gray
      },
      /pgf/number format/.cd,
      use comma,
      set thousands separator={},
   } 
} 

\begin{document} 

 \begin{tikzpicture} 
    \begin{axis}[01_temp]
       \addplot[black, thin, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \addlegendentry{} 
       \addplot[black, very thick, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70101}
       \addplot[red, thin, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \addlegendentry{} 
         \addplot[red, very thick, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70201}
       \addplot[green, thin, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \addlegendentry{}
         \addplot[green, very thick, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70301}
         \addplot[blue, thin, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \addlegendentry{}
         \addplot[blue, very thick, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70401}
         \addplot[pink, thin, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \addlegendentry{}
         \addplot[pink, very thick, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \addlegendentry{K 70501}
    \end{axis} 
 \end{tikzpicture} 
\end{document}

Unfortunately, I didn't manage it to change the vertical spacing between those sample lines in the legend. – I tried a little bit with "yshift" but as I had to notice, this seems to be not the right way …

Maybe someone could help me … that would be really nice! =)

Thank you in advance,
eniem

Best Answer

You can label each plot and then set the legend as a tikz matrix outside the axis environment. The two related legendimages can be placed in the same matrix cell using \raisebox and \llap.

\documentclass[tikz,border=5pt]{standalone}
\usepackage[ngerman]{babel}
\usepackage[utf8]{inputenc}
\usepackage[T1]{fontenc}
\usepackage{pgfplots}
\usetikzlibrary{matrix}
\usepackage{amsmath} 
% Settings for tikz-Image
\pgfplotsset{compat=1.10, 
   01_temp/.style={
      %title=Knotentemperaturen,
      xmin=0,
      xmax=200,
      xlabel={Zeit $t$ [h]},
      x tick style={
        color=black,
        thin
      },      
      ymin=15,
      ymax=50,
      ylabel={Knotentemperatur $T$ [$^\circ$C]},
      y tick style={
        color=black,
        thin
      },
      height=9cm,
      width=15cm,
      grid=major,
      grid style={
         solid,
         ultra thin,
         gray
      },
      /pgf/number format/.cd,
      use comma,
      set thousands separator={},
   }
} 

\newcommand\refentry[1]{% positions two related legendimages in one cell
  \raisebox{1.5pt}{\ref{plot:#1a}}\llap{\raisebox{-1pt}{\ref{plot:#1b}}}%
}

\begin{document} 

 \begin{tikzpicture} 
    \begin{axis}[01_temp]
       \addplot[black, thin, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \label{plot:K70101a} 
       \addplot[black, very thick, solid] table [x index=0,y index=1, skip first n=8] {01_temp.dat}; 
       \label{plot:K70101b}
       \addplot[red, thin, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \label{plot:K70201a} 
       \addplot[red, very thick, dotted] table [x index=0,y index=2, skip first n=8] {01_temp.dat}; 
       \label{plot:K70201b}
       \addplot[green, thin, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \label{plot:K70301a}
       \addplot[green, very thick, dashed] table [x index=0,y index=3, skip first n=8] {01_temp.dat}; 
       \label{plot:K70301b}
       \addplot[blue, thin, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \label{plot:K70401a}
       \addplot[blue, very thick, dashed] table [x index=0,y index=4, skip first n=8] {01_temp.dat}; 
       \label{plot:K70401b}
       \addplot[pink, thin, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \label{plot:K70501a}
       \addplot[pink, very thick, dashed] table [x index=0,y index=5, skip first n=8] {01_temp.dat}; 
       \label{plot:K70501b}
    \end{axis}
% Legend
    \matrix[
        matrix of nodes,
        anchor=north east,
        inner sep=0.2em,
        nodes={font=\scriptsize},
      ] at([yshift=-1ex]current axis.north east)
      {
        \refentry{K70101}& K70101&[2pt]\refentry{K70201}& K70201&[2pt]\refentry{K70301}& K70301\\
        \refentry{K70401}& K70401&[2pt]\refentry{K70501}& K70501\\};
\end{tikzpicture}
\end{document}

Run twice to get

enter image description here

If the background of the legend should be colored use fill=<color> as a matrixoption.

% Legend
    \matrix[
        matrix of nodes,
        anchor=north east,
        inner sep=0.2em,
        fill=gray!20,% background color of the legend
        nodes={font=\scriptsize},
      ] at([xshift=-1ex,yshift=-1ex]current axis.north east)
      {
        \refentry{K70101}& K70101&[2pt]\refentry{K70201}& K70201&[2pt]\refentry{K70301}& K70301\\
        \refentry{K70401}& K70401&[2pt]\refentry{K70501}& K70501\\};

enter image description here

Related Solutions

[Tex/LaTex] pgfplots particular legend

To add a legend, you just have to call \addlegendentry{label text} at least once, or set the entry/entries in some other way (using \legend{<list of label texts>} or the legend entries={<list of label texts>} key). By default, the label texts are associated with the plots in the sequence they were created.

If, in your example, you add the command \addlegendentry{$f(x)$ after your \addplot command (or before, it doesn't technically matter, but it's more semantic to add it after). Then you will get

Note that a line is shown in the legend, even though it doesn't appear in the plot. That's because you created a line plot, and the legend image doesn't take your filling into account. If you use \addplot [area style] ..., the legend will look like

To add more lines to the legend without associated plots, you can call \addlegendimage{empty legend} and then \addlegendentry as usual. To have the label texts left aligned, you can call legend cell align=left.

\addlegendimage{empty legend}\addlegendentry{$\mu=0$}
\addlegendimage{empty legend}\addlegendentry{$\sigma^2=0.2$}

Here's the complete code:

\documentclass{article}
\usepackage{pgfplots}
\begin{document}
\pgfplotsset{width=10cm}
\begin{tikzpicture}
\makeatletter \newcommand{\pgfplotsdrawaxis}{\pgfplots@draw@axis} \makeatother
\pgfplotsset{axis line on top/.style={
 axis on top=false,
 after end axis/.append code={
    \pgfplotsset{axis line style=opaque,
      ticklabel style=opaque,
      tick style=opaque,
      grid=none}
    \pgfplotsdrawaxis}
  }
}
\pgfdeclarehorizontalshading{stefan}{50bp}{
color(0.00000000000000bp)=(violet);
color(8.33333333333333bp)=(blue);
color(16.66666666666670bp)=(cyan);
color(25.00000000000000bp)=(green);
color(33.33333333333330bp)=(yellow);
color(41.66666666666670bp)=(orange);
color(50.00000000000000bp)=(red)
}
\begin{axis}[
grid=major,
xmin=-0.3, xmax=0.3,
ymin=0, ymax=1,
point meta min={-0.3},
point meta max={0.3},
axis line on top,
tick style={thin,black},
xticklabel style={text height=1.5ex},
xticklabels={%
$-0.4$,
$-0.3$,
$-0.2$,
$-0.1$,
$\mu$,
$0.1$,
$0.2$,
$0.3$,
$0.4$},
extra x ticks={-0.4,-0.3,...,0.4},
extra x tick labels={%
$-0.4$,
$-0.3$,
$-0.2$,
$-0.1$,
$\mu$,
$0.1$,
$0.2$,
$0.3$,
$0.4$},
extra x tick style={
    xticklabel pos=right,
    xticklabel style={text depth=0pt}
},
extra y ticks={0,0.2,...,1},
extra y tick style={
    yticklabel pos=right
},
minor tick num=4,
xlabel=\large $x$,
ylabel=\large $f_{\mu,\sigma^2}(x)$,
colorbar horizontal,
colorbar style={
    xticklabels={%
        $-0.4$,
        $-0.3$,
        $-0.2$,
        $-0.1$,
        $\mu$,
        $0.1$,
        $0.2$,
        $0.3$,
        $0.4$
    },
    xticklabel style={text height=1.5ex}
},
no markers,
colormap={new}{color(0cm)=(violet);color(1cm)=(blue);color(2cm)=(cyan);color(3cm)=(green);color(4cm)=(yellow);color(5cm)=(orange);color(6cm)=(red)},
after end axis/.append code={
    \draw ({rel axis cs:0,0}-|{axis cs:0,0}) -- ({rel axis cs:0,1}-|{axis cs:0,0});
},
legend cell align=left]
\addplot[area style] gnuplot[
    shading=stefan,
    draw=none,
    shader=interp,
    id=DoG,
    samples=1000,
    domain=-0.3:0.3,
    y domain=0:1
]{((1/(sqrt(2*pi*0.00570275999999999)))*exp(-(x-0)**2/(2*0.00570275999999999))/6.603550926)}\closedcycle;
\addlegendentry{$f(x)$}
\addlegendimage{empty legend}\addlegendentry{$\mu=0$}
\addlegendimage{empty legend}\addlegendentry{$\sigma^2=0.2$}
\end{axis}
\end{tikzpicture}
\end{document}

[Tex/LaTex] plotting two time series with bounds

This happens because PGFPlots only uses one "stack" per axis: You're stacking the second confidence interval on top of the first. The easiest way to fix this is probably to use the approach described in "Is there an easy way of using line thickness as error indicator in a plot?": After plotting the first confidence interval, stack the upper bound on top again, using stack dir=minus. That way, the stack will be reset to zero, and you can draw the second confidence interval in the same fashion as the first:

\documentclass{standalone}
\usepackage{pgfplots, tikz}

\usepackage{pgfplotstable}

\pgfplotstableread{
temps   y_h y_h__inf    y_h__sup    y_f y_f__inf    y_f__sup    

1   0.237340    0.135170    0.339511    0.237653    0.135482    0.339823    
2   0.561320    0.422007    0.700633    0.165871    0.026558    0.305184    
3   0.694760    0.534205    0.855314    0.074856    -0.085698   0.235411    
4   0.728306    0.560179    0.896432    0.003361    -0.164765   0.171487    
5   0.711710    0.544944    0.878477    -0.044582   -0.211349   0.122184    
6   0.671241    0.511191    0.831291    -0.073347   -0.233397   0.086703    
7   0.621177    0.471219    0.771135    -0.088418   -0.238376   0.061540    
8   0.569354    0.431826    0.706882    -0.094382   -0.231910   0.043146    
9   0.519973    0.396571    0.643376    -0.094619   -0.218022   0.028783    
10  0.475121    0.366990    0.583251    -0.091467   -0.199598   0.016664    
}{\table}

\begin{document}
\begin{tikzpicture}

    \begin{axis}
    % y_h confidence interval
    \addplot [stack plots=y, fill=none, draw=none, forget plot]   table [x=temps, y=y_h__inf]   {\table} \closedcycle;
    \addplot [stack plots=y, fill=gray!50, opacity=0.4, draw opacity=0, area legend]   table [x=temps, y expr=\thisrow{y_h__sup}-\thisrow{y_h__inf}]   {\table} \closedcycle;
    % subtract the upper bound so our stack is back at zero
    \addplot [stack plots=y, stack dir=minus, forget plot, draw=none] table [x=temps, y=y_h__sup] {\table};

    % y_f confidence interval
    \addplot [stack plots=y, fill=none, draw=none, forget plot]   table [x=temps, y=y_f__inf]   {\table} \closedcycle;
    \addplot [stack plots=y, fill=gray!50, opacity=0.4, draw opacity=0, area legend]   table [x=temps, y expr=\thisrow{y_f__sup}-\thisrow{y_f__inf}]   {\table} \closedcycle;

    % the line plots (y_h and y_f)    
    \addplot [stack plots=false, very thick,smooth,blue]  table [x=temps, y=y_h]   {\table};
    \addplot [stack plots=false, very thick,smooth,blue]  table [x=temps, y=y_f]   {\table};
    \end{axis}

\end{tikzpicture}
\end{document}

Best Answer

Related Solutions

[Tex/LaTex] pgfplots particular legend

[Tex/LaTex] plotting two time series with bounds

Related Question