Pgfplots: Discrete instead of connected values

Question

I have a table with experimental data. The table has two columns separated by a semicolon. The first column is the value for the x-axis and the second one is the y-axis value.
If I use the general addplot table {} command, pgfplots draws a plot where it connects the points together. As the data points in question are discrete values, there should not be a connection between them, but rather should every point be drawn independently.
What's the correct approach for this using pgfplots?
Output generated by the addplot table {} command

documentclass{article}

usepackage{pgfplots}
pgfplotsset{compat = 1.17}
usepackage{siunitx}

begin{filecontents}{table.csv}
    57.7200;403.7083
    81.1200;359.9954
    83.7600;2673.1575
    84.4800;201.8542
    95.7600;359.9954
    104.6400;403.7083
    104.8800;2673.1575
    105.0000;5146.4829
    105.1200;11393.4863
    105.2400;5706.3101
    105.9600;201.8542
    106.9200;4277.3057
    107.2800;359.9954
    107.4000;682.9842
    107.8800;2673.1575
    108.0000;403.7083
    108.3600;2673.1575
    108.6000;201.8542
    109.2000;47.6190
    109.4400;3883.1460
    109.9200;4081.7891
    110.0400;1113.2275
end{filecontents}

begin{document}

begin{tikzpicture}
    
    begin{axis}[
        xlabel = {textit{m/z}},
        ylabel = {Int. [si{cps}]},
    ]
        addplot+[mark = none, red] table[
            col sep = semicolon
        ] {table.csv};
    end{axis}

end{tikzpicture}

end{document}

Expected output

This output was generated by a little "hack". To have the correct axis, I drew the plot as before but changed the color to red!0 so that it is transparent. Then, for each table entry, I manually drew the lines with the draw command. It goes without mentioning, that this is not optimal, but at least it illustrates the desired output.
documentclass{article}

usepackage{pgfplots}
pgfplotsset{compat = 1.17}
usepackage{siunitx}

begin{filecontents}{table.csv}
    57.7200;403.7083
    81.1200;359.9954
    83.7600;2673.1575
    84.4800;201.8542
    95.7600;359.9954
    104.6400;403.7083
    104.8800;2673.1575
    105.0000;5146.4829
    105.1200;11393.4863
    105.2400;5706.3101
    105.9600;201.8542
    106.9200;4277.3057
    107.2800;359.9954
    107.4000;682.9842
    107.8800;2673.1575
    108.0000;403.7083
    108.3600;2673.1575
    108.6000;201.8542
    109.2000;47.6190
    109.4400;3883.1460
    109.9200;4081.7891
    110.0400;1113.2275
end{filecontents}

begin{document}

begin{tikzpicture}
    
    begin{axis}[
        xlabel = {textit{m/z}},
        ylabel = {Int. [si{cps}]},
    ]
        addplot+[mark = none, red!0] table[
            col sep = semicolon
        ] {table.csv};
        draw[red] (57.7200,0) -- (57.7200,403.7083);
        draw[red] (81.1200,0) -- (81.1200,359.9954);
        draw[red] (83.7600,0) -- (83.7600,2673.1575);
        draw[red] (84.4800,0) -- (84.4800,201.8542);
        draw[red] (95.7600,0) -- (95.7600,359.9954);
        draw[red] (104.6400,0) -- (104.6400,403.7083);
        draw[red] (104.8800,0) -- (104.8800,2673.1575);
        draw[red] (105.0000,0) -- (105.0000,5146.4829);
        draw[red] (105.1200,0) -- (105.1200,11393.4863);
        draw[red] (105.2400,0) -- (105.2400,5706.3101);
        draw[red] (105.9600,0) -- (105.9600,201.8542);
        draw[red] (106.9200,0) -- (106.9200,4277.3057);
        draw[red] (107.2800,0) -- (107.2800,359.9954);
        draw[red] (107.4000,0) -- (107.4000,682.9842);
        draw[red] (107.8800,0) -- (107.8800,2673.1575);
        draw[red] (108.0000,0) -- (108.0000,403.7083);
        draw[red] (108.3600,0) -- (108.3600,2673.1575);
        draw[red] (108.6000,0) -- (108.6000,201.8542);
        draw[red] (109.2000,0) -- (109.2000,47.6190);
        draw[red] (109.4400,0) -- (109.4400,3883.1460);
        draw[red] (109.9200,0) -- (109.9200,4081.7891);
        draw[red] (110.0400,0) -- (110.0400,1113.2275);
    end{axis}

end{tikzpicture}

end{document}

user227987 · Accepted Answer

I think you are looking for ycomb.
documentclass{article}

usepackage{pgfplots}
pgfplotsset{compat = 1.17}
usepackage{siunitx}

begin{filecontents}[overwrite]{table.csv}
    57.7200;403.7083
    81.1200;359.9954
    83.7600;2673.1575
    84.4800;201.8542
    95.7600;359.9954
    104.6400;403.7083
    104.8800;2673.1575
    105.0000;5146.4829
    105.1200;11393.4863
    105.2400;5706.3101
    105.9600;201.8542
    106.9200;4277.3057
    107.2800;359.9954
    107.4000;682.9842
    107.8800;2673.1575
    108.0000;403.7083
    108.3600;2673.1575
    108.6000;201.8542
    109.2000;47.6190
    109.4400;3883.1460
    109.9200;4081.7891
    110.0400;1113.2275
end{filecontents}

begin{document}

begin{tikzpicture}
    
    begin{axis}[
        xlabel = {textit{m/z}},
        ylabel = {Int. [si{cps}]},
    ]
        addplot+[mark = none, red,ycomb] table[
            col sep = semicolon
        ] {table.csv};
    end{axis}

end{tikzpicture}

end{document}

frougon · Answer

As shown in anonymous' answer, ycomb can do the job. Another way is to use ybar. This plots each data point as a vertical bar (for horizontal bars, you would use xbar). The bar width can be customized using bar width, as in:
begin{axis}[..., ybar, bar width=0.4pt]
    addplot table[draw=none] {table.csv};
end{axis}

As you found out yourself, draw=none allows one to get the precise specified width for the bars, otherwise they are drawn slightly larger due to the draw operation.
Full example:
documentclass{article}
usepackage{pgfplots}
pgfplotsset{compat=1.17}
usepackage{siunitx}

begin{filecontents}{table.csv}
    57.7200;403.7083
    81.1200;359.9954
    83.7600;2673.1575
    84.4800;201.8542
    95.7600;359.9954
    104.6400;403.7083
    104.8800;2673.1575
    105.0000;5146.4829
    105.1200;11393.4863
    105.2400;5706.3101
    105.9600;201.8542
    106.9200;4277.3057
    107.2800;359.9954
    107.4000;682.9842
    107.8800;2673.1575
    108.0000;403.7083
    108.3600;2673.1575
    108.6000;201.8542
    109.2000;47.6190
    109.4400;3883.1460
    109.9200;4081.7891
    110.0400;1113.2275
end{filecontents}

begin{document}

begin{tikzpicture}
    begin{axis}[
        xlabel = {textit{m/z}},
        ylabel = {Int. [si{cps}]},
        ybar, bar width = 0.4pt,
    ]
        addplot+[draw=none, fill=red] table[col sep = semicolon] {table.csv};
    end{axis}
end{tikzpicture}

end{document}

If you want the bar width to be equal to the current PGF line width in the tikzpicture, you can use bar width/.expanded = thepgflinewidth:
documentclass{article}
usepackage{pgfplots}
pgfplotsset{compat=1.17}
usepackage{siunitx}

begin{filecontents}{table.csv}
    57.7200;403.7083
    81.1200;359.9954
    83.7600;2673.1575
    84.4800;201.8542
    95.7600;359.9954
    104.6400;403.7083
    104.8800;2673.1575
    105.0000;5146.4829
    105.1200;11393.4863
    105.2400;5706.3101
    105.9600;201.8542
    106.9200;4277.3057
    107.2800;359.9954
    107.4000;682.9842
    107.8800;2673.1575
    108.0000;403.7083
    108.3600;2673.1575
    108.6000;201.8542
    109.2000;47.6190
    109.4400;3883.1460
    109.9200;4081.7891
    110.0400;1113.2275
end{filecontents}

begin{document}

begin{tikzpicture}[line width=2pt]
    begin{axis}[
        xlabel = {textit{m/z}},
        ylabel = {Int. [si{cps}]},
        ybar, bar width/.expanded = thepgflinewidth,
    ]
        addplot+[draw=none, fill=red] table[col sep = semicolon] {table.csv};
    end{axis}
end{tikzpicture}

end{document}

Pgfplots: Discrete instead of connected values

Output generated by the `addplot table {}` command

Expected output

2 Answers

Add your own answers!

Ask a Question