I'm trying to create a table that looks like this .

Currently I'm doing this ...

\begin{table}[ht!]
\centering
\begin{tabular}{|l|l|l|l|l|l|l|l|l|l|}
\hline
\multicolumn{1}{|c|}{\textbf{Patch}} & \multicolumn{3}{c|}{\textbf{ACES RGB}} & \multicolumn{3}{c|}{\textbf{Display RGB}} & \multicolumn{3}{c|}{\textbf{Display xyY}} \\ \hline
\textbf{N1} & 1.8233 & 1.8233 & 1.8233 & 0.9243 & 0.8651 & 0.9013 & 0.3217 & 0.3377 & 34.4858 \\ \hline
\textbf{N2} & 0.2753 & 0.2753 & 0.2753 & 0.5383 & 0.5038 & 0.5249 & 0.3217 & 0.3377 & 8.4552  \\ \hline
\textbf{N3} & 0.0898 & 0.0898 & 0.0898 & 0.2804 & 0.2625 & 0.2734 & 0.3217 & 0.3377 & 1.5514  \\ \hline
\textbf{R}  & 0.4689 & 0.1193 & 0.0417 & 0.8046 & 0.2227 & 0.1795 & 0.6413 & 0.3307 & 6.4488  \\ \hline
\textbf{G}  & 0.339  & 0.8068 & 0.0936 & 0.4335 & 0.8036 & 0.2434 & 0.3046 & 0.624  & 20.8422 \\ \hline
\textbf{B}  & 0.2162 & 0.133  & 0.8711 & 0.1707 & 0.1503 & 0.8215 & 0.1562 & 0.0692 & 2.3365  \\ \hline
\textbf{C}  & 0.5187 & 0.9138 & 1.0432 & 0.4332 & 0.8028 & 0.8406 & 0.2269 & 0.3404 & 22.8164 \\ \hline
\textbf{M}  & 0.58   & 0.2096 & 0.9086 & 0.808  & 0.2134 & 0.8294 & 0.333  & 0.1596 & 8.4349  \\ \hline
\textbf{Y}  & 0.8237 & 0.9378 & 0.0855 & 0.8654 & 0.8096 & 0.2487 & 0.4338 & 0.5187 & 26.9923 \\ \hline
\end{tabular}
\end{table}


I have many tables to create that all contain the same number and titled rows and columns (e.g. the bold stuff in the example above is the same for all tables). Unfortunately, building each table is tedious.

I'm trying to figure out how I can use csvsimple to load the variable values in the table from a .csv file, but I can't seem to figure out how to the multicolumn and header rows to work.

I'm thinking each csv file would be formatted like this:

1.8233,1.8233,1.8233,0.9243,0.8651,0.9013,0.3217,0.3377,34.4858
0.2753,0.2753,0.2753,0.5383,0.5038,0.5249,0.3217,0.3377,8.4552
0.0898,0.0898,0.0898,0.2804,0.2625,0.2734,0.3217,0.3377,1.5514
0.4689,0.1193,0.0417,0.8046,0.2227,0.1795,0.6413,0.3307,6.4488
0.339,0.8068,0.0936,0.4335,0.8036,0.2434,0.3046,0.624,20.8422
0.2162,0.133,0.8711,0.1707,0.1503,0.8215,0.1562,0.0692,2.3365
0.5187,0.9138,1.0432,0.4332,0.8028,0.8406,0.2269,0.3404,22.8164
0.58,0.2096,0.9086,0.808,0.2134,0.8294,0.333,0.1596,8.4349
0.8237,0.9378,0.0855,0.8654,0.8096,0.2487,0.4338,0.5187,26.9923

• My personal approach to this kind of problems is to write a small python¹ script that transforms the cvs file into a ready-made LaTeX table and then \input that table. ¹Of course, you can use a different scripting language when you know it better. – jknappen Nov 29 '17 at 11:02
• @jknappen sure I could do this but part of what I’m trying to avoid is large sections of code being dedicated to just building a table. I was planning on ultimately making a macro that I could reuse. – agf1997 Nov 29 '17 at 16:44

The first solution is with booktabs and without vertical rules (I think the table looks better this way); the second one is like the OP's image.

Edit: I've added a third example which doesn't use siunitx, to show the difference to the OP.

\documentclass{article}
\usepackage{siunitx}
\usepackage{array}
\usepackage{booktabs}
\usepackage{csvsimple}
\usepackage{caption}

\begin{filecontents*}{mydata.csv}
1.8233,1.8233,1.8233,0.9243,0.8651,0.9013,0.3217,0.3377,34.4858
0.2753,0.2753,0.2753,0.5383,0.5038,0.5249,0.3217,0.3377,8.4552
0.0898,0.0898,0.0898,0.2804,0.2625,0.2734,0.3217,0.3377,1.5514
0.4689,0.1193,0.0417,0.8046,0.2227,0.1795,0.6413,0.3307,6.4488
0.339,0.8068,0.0936,0.4335,0.8036,0.2434,0.3046,0.624,20.8422
0.2162,0.133,0.8711,0.1707,0.1503,0.8215,0.1562,0.0692,2.3365
0.5187,0.9138,1.0432,0.4332,0.8028,0.8406,0.2269,0.3404,22.8164
0.58,0.2096,0.9086,0.808,0.2134,0.8294,0.333,0.1596,8.4349
0.8237,0.9378,0.0855,0.8654,0.8096,0.2487,0.4338,0.5187,26.9923
\end{filecontents*}

\begin{document}
\begin{table}[ht!]
\caption{The best}
\centering\tiny\renewcommand*{\arraystretch}{1.4}
\begin{tabular}{>{\bfseries\arraybackslash}c}
\toprule
Patch\\\cmidrule(lr){1-1}
N1\\ N2\\ N3\\ R\\ G\\ B\\ C\\ M\\ Y\\
\bottomrule
\end{tabular}%
,
tabular={c@{}*8{S[table-format=1.4]}S[table-format=2.4]},
table head=\toprule & \multicolumn{3}{c}{\textbf{ACES RGB}} & \multicolumn{3}{c}{\textbf{Display RGB}} & \multicolumn{3}{c}{\textbf{Display xyY}}\\
\cmidrule(lr){1-4} \cmidrule(lr){5-7} \cmidrule(lr){8-10}, command=&\csvcoli & \csvcolii & \csvcoliii & \csvcoliv & \csvcolv & \csvcolvi & \csvcolvii & \csvcolviii & \csvcolix, table foot=\bottomrule}
\end{table}

\begin{table}[ht!]
\caption{The not-so-beautiful}
\centering\tiny\renewcommand*{\arraystretch}{2}
\begin{tabular}{|>{\bfseries\arraybackslash}c|}
\hline
Patch\\\hline
N1\\\hline N2\\\hline N3\\\hline
R\\\hline G\\\hline B\\\hline
C\\\hline M\\\hline Y\\\hline
\end{tabular}%
,
tabular={c@{}*8{S[table-format=1.4]|}S[table-format=2.4]|},
table head=\hline & \multicolumn{3}{c|}{\textbf{ACES RGB}} & \multicolumn{3}{c|}{\textbf{Display RGB}} & \multicolumn{3}{c|}{\textbf{Display xyY}}\\
\hline,
command=&\csvcoli & \csvcolii & \csvcoliii & \csvcoliv & \csvcolv & \csvcolvi & \csvcolvii & \csvcolviii & \csvcolix,
late after line=\\\hline}
\end{table}

\begin{table}[ht!]
\caption{The ugly}
\centering\tiny\renewcommand*{\arraystretch}{2}
\begin{tabular}{|>{\bfseries\arraybackslash}c|}
\hline
Patch\\\hline
N1\\\hline N2\\\hline N3\\\hline
R\\\hline G\\\hline B\\\hline
C\\\hline M\\\hline Y\\\hline
\end{tabular}%
tabular={*9{l|}},
table head=\hline \multicolumn{3}{c|}{\textbf{ACES RGB}} & \multicolumn{3}{c|}{\textbf{Display RGB}} & \multicolumn{3}{c|}{\textbf{Display xyY}}\\
\hline,
command=\csvcoli & \csvcolii & \csvcoliii & \csvcoliv & \csvcolv & \csvcolvi & \csvcolvii & \csvcolviii & \csvcolix,
late after line=\\\hline}
\end{table}
\end{document}


• why use before reading=\centering\sisetup{table-number-alignment=center instead of just before reading=\centering which doesn't require siunitx? – agf1997 Nov 29 '17 at 22:11
• @agf1997 \sisetup{table-number-alignment=center} isn't necessary, but I used siunitx to better aling the columns, see column types S[table-format=...]. – CarLaTeX Nov 29 '17 at 22:20
• @agf1997 Please look at my renewed answer, the "ugly" example is without siunitx, please note the alignment of the last column. – CarLaTeX Nov 30 '17 at 5:32
• yes this is very helpful. Thank you so much – agf1997 Nov 30 '17 at 17:56
• @agf1997 You're welcome, thank you for accepting my answer! – CarLaTeX Nov 30 '17 at 18:00