I've tried to use pgfplots to create a scatterplot with marginal histograms like it can be seen here:
Creating the histograms and the scatterplot alone works, but putting it together results in a desaster. Does pgfplots supports this kind of graph?
[Tex/LaTex] Scatterplot with marginal histograms using pgfplots
pgfplots
Related Solutions
You need to add the ybar option to the axis options, instead of to each plot individually. Note that the author of pgfplots recommends not to use the newest compat option, and also I think that loading the pgfplotslibrary units is redundant if you also load siunitx. Finally, pgfplots already loads tikz so no need for duplication there either :-)
\documentclass{article}
\usepackage{siunitx}
\usepackage{pgfplots}
\pgfplotsset{compat=1.12}
\sisetup{
round-mode = places,
round-precision = 2,
}
\begin{document}
\begin{figure}[h!]
\begin{center}
\begin{tikzpicture}
\begin{axis}[
width=15cm, height=8cm,
grid=major,
grid style={dashed,gray!30},
xlabel=String length,
ylabel=Performance ops/ms,
legend style={at={(0.5,-0.2)},anchor=north},
xmode=log,
log ticks with fixed point,
xtick=data,
ybar, %added here
]
\addplot[fill] table[x="length",y="Score", col sep=comma] {jmh-result1.csv};
\addplot[green,fill] table[x="length",y="Score", col sep=comma] {jmh-result2.csv};
\end{axis}
\end{tikzpicture}
\caption{Performance Comparison Histogram}
\end{center}
\end{figure}
\end{document}
You can use it as follows: Don't use macro names starting with \the. That is a special case for TeX and might lead to mistakes that are very difficult to debug.
\documentclass{article}
\usepackage{pgfplotstable}
\begin{document}
\begin{tikzpicture}
\pgfplotstableread[col sep=comma,header=false]{
12.3458,0.709423,0.018174,10.3177,0.031258,0.360285,0.071809,0
13.3458,0.709423,0.018174,10.3177,0.031258,0.360285,0.171809,0
7.88918,0.037782,0.010597,13.0123,0.027078,0.345659,0.070872,1
8.88918,0.037782,0.010597,13.0123,0.027078,0.345659,0.170872,1
3.29679,0.175776,0.012142,18.2475,0.031448,0.292123,0.141521,2
4.29679,0.175776,0.012142,18.2475,0.031448,0.292123,0.241521,2
3.94161,0.204657,0.002334,2.09774,0.011567,0.278266,0.113811,3
4.94161,0.204657,0.002334,2.09774,0.011567,0.278266,0.213811,3
}\mydata
\begin{axis}[width=7cm,height=7cm]
\addplot+[scatter, only marks,
scatter/classes={0={mark=square*,green},
1={mark=triangle*,black},
2={mark=o,blue},
3={mark=x,red}
},
scatter src=explicit symbolic
] table[x index=0,y index=6,meta index=7] \mydata;
\end{axis}
\end{tikzpicture}
\end{document}
Best Answer
PGFPlots can calculate histograms. By using three
axisenvironments positioned relative to each other, you can get the desired output:random.dat: