Tables – Understanding the Table-Format Role in Tabularx

Question

As a follow up to my question here

I tried to change the 5-column table in the question above to a 4 column table. Now, I want to change it to a 3 column table.

Here is my 4 column table:

\begin{table*}
\centering
\caption{ Significance test between top performing models.}
\label{table:sign_tok}
\begin{tabular}{@{}lS[table-format=4] *{2}{S[table-format=2.1]@{\;(}S[table-format=2.1(2)]@{)\;}l} @{}}
  \toprule
 {\thead{Source}}& {\thead{Target}} & \multicolumn{3}{l}{\thead[l]{BLEU $(\mu \pm 0.95 \%)$}}  & \multicolumn{3}{l}{\thead[l]{CHRF2++ $(\mu \pm 0.95 \%)$}}   \\ 
  \midrule
TokA  & TokB   & 29.7 & 29.7(8) &      & 53.1 & 53.1(7)  \\
TokC  & TokW   & 31.6 & 31.6(8) & (p=0.0010)* & 52.2 & 52.2(7) & (p = 0.0010)* \\

  
 \bottomrule
\end{tabular}
\end{table*} 

This is how my 4-column table looks:

How can I remove the first column of the above table to change it to a 3 column table

I believe changing this line of the latex code will do the job, but I don't quite understand what it does if someone could explain it to me.

\begin{tabular}{@{}lS[table-format=4] *{2}{S[table-format=2.1]@{\;(}S[table-format=2.1(2)]@{)\;}l} @{}}

Answer 1

See, if the following solution works for you:

\documentclass[twocolumn]{article}
\usepackage{booktabs, makecell, multirow}
\renewcommand\theadfont{\normalsize\bfseries}
\renewcommand\theadgape{}
\usepackage{siunitx}
\sisetup{
output-open-uncertainty = (,
output-close-uncertainty = ),
uncertainty-mode=separate
}%

\begin{document}
    \begin{table*}
    \centering
    \caption{ Significance test between top performing models.}
    \label{table:sign_tok}
\begin{tabular}{@{}lc 
               *{2}{S[table-format=2.1] 
                     @{\;(} 
                    S[table-format=2.1(2)] 
                     @{\,)\quad } 
                    l} @{}}
    \toprule
\multicolumn{2}{c}{\thead{Segmentation}}
        & \multicolumn{6}{c}{}  \\
        \cmidrule(r){1-2}
 \thead{Source}
    & \thead{Target}
        & \multicolumn{3}{l}{\multirow{-2.2}*{\thead[l]{BLEU $(\mu \pm 0.95 \%)$}}}
                    & \multicolumn{3}{l}{\multirow{-2.2}*{\thead[l]{CHRF2++ $(\mu \pm 0.95 \%)$}}}
                            \\
    \midrule
SPM  & SPM   & 29.7 & 29.7(8)   &             & 53.1 & 53.1(7) &                \\
ATB  & RAW   & 31.6 & 31.6(8)   & (p=0.0010)* & 52.2 & 52.2(7) & (p = 0.0010)*  \\
D3   & RAW   & 32.9 & 32.9(10)  & (p=0.0010)* & 53.4 & 53.4(6) & (p = 0.0909)   \\
SPM  & RAW   & 29.7 & 29.7(8)   & (p=0.4256)  & 50.4 & 50.4(7) & (p = 0.0010)*  \\
    \bottomrule
\end{tabular}
    \end{table*} 
\end{document} 

Addendum To have only three columns (well, I would rather have four):

\documentclass[twocolumn]{article}
\usepackage{booktabs, makecell, multirow}
\renewcommand\theadfont{\normalsize\bfseries}
\renewcommand\theadgape{}
\usepackage{siunitx}
\sisetup{
uncertainty-mode=separate
}%

\begin{document}
    \begin{table*}
    \centering
    \caption{ Significance test between top performing models.}
    \label{table:sign_tok}
\begin{tabular}{@{} lll @{}}
    \toprule
\thead{Segmentation of\\ Source and Target}
        & \thead[l]{BLEU $(\mu \pm 0.95 \%)$}
                    & \thead[l]{CHRF2++ $(\mu \pm 0.95 \%)$}   \\
    \midrule
SPM SPM   & 29.7 (\num{29.7(8)})              & 53.1 (\num{53.1(7)})                \\
ATB RAW   & 31.6 (\num{31.6(8)})  (p=0.0010)* & 52.2 (\num{52.2(7)}) (p = 0.0010)*  \\
D3 RAW    & 32.9 (\num{32.9(10)}) (p=0.0010)* & 53.4 (\num{53.4(6)}) (p = 0.0909)   \\
SPM RAW   & 29.7 (\num{29.7(8)})  (p=0.4256)  & 50.4 (\num{50.4(7)}) (p = 0.0010)*  \\
    \bottomrule
\end{tabular}
    \end{table*} 
\end{document} 

With four columns is much nicer (to my opinion):

\documentclass[twocolumn]{article}
\usepackage{booktabs, makecell, multirow}
\renewcommand\theadfont{\normalsize\bfseries}
\renewcommand\theadgape{}
\usepackage{siunitx}
\sisetup{
uncertainty-mode=separate
}%

\begin{document}
    \begin{table*}
    \centering
    \caption{ Significance test between top performing models.}
    \label{table:sign_tok}
\begin{tabular}{@{} ccll @{}}
    \toprule
\multicolumn{2}{c}{\thead{Segmentation}}
        & \multirow{2.2}{*}{\thead[l]{BLEU $(\mu \pm 0.95 \%)$}}
            & \multirow{2.2}{*}{\thead[l]{CHRF2++ $(\mu \pm 0.95 \%)$}}   \\
    \cmidrule(r){1-2}
\thead{Source}
    & \thead{Target}
        &   &   \\
     \midrule
SPM & SPM   & 29.7 (\num{29.7(8)})              & 53.1 (\num{53.1(7)})                \\
ATB & RAW   & 31.6 (\num{31.6(8)})  (p=0.0010)* & 52.2 (\num{52.2(7)}) (p = 0.0010)*  \\
D3  & RAW   & 32.9 (\num{32.9(10)}) (p=0.0010)* & 53.4 (\num{53.4(6)}) (p = 0.0909)   \\
SPM & RAW   & 29.7 (\num{29.7(8)})  (p=0.4256)  & 50.4 (\num{50.4(7)}) (p = 0.0010)*  \\
    \bottomrule
\end{tabular}
    \end{table*} 
\end{document}