Reduce the column padding in tabular environment

Question

I've tried nearly everything with no success and I don't know why... I have a tablular with 9 columns and the only solution I found for fitting it into textwidth is using adjustbox, which I don't like because it scale the fonts it seems. So the numbers on my table can't barely be seen:

I tried several other methods and the only one I found to "work" is to set manually the column padding instead of using adjustbox by adding to each one: @{\hspace{0.5em}}:

But then I have those unaligned vertical lines and I don't know how to deal with it... It's been like 4 hours since I started to solve this problem and nothing worked at all.

All I want is to adjust the table to the textwidth and make the number bigger by reducing the padding in each column. Can anyone help me a little bit? Thanks!

\begin{table}
\centering
\begin{adjustbox}{max width=\textwidth}
\begin{tabular}{|c|c|c|c|c|c|c|c|c|}
   \multicolumn{1}{c}{} &  \multicolumn{4}{c|}{Conjunto de datos original}  & \multicolumn{4}{c}{Conjunto de datos reducido} \\
    \cline{2-9}
   \multicolumn{1}{c|}{} & Filtro I & Filtro R & Filtro V & Filtros I,R,V & Filtro I & Filtro R & Filtro V & Filtros I,R,V \\

    \hline

   \begin{tabular}[c]{@{}c@{}}$q=0.62$\\ \citep{2007ApJ...662..564B} \end{tabular}  & $43.2_{-4.7}^{+4.2}$ & $41.9_{-3.4}^{+3.6}$ & $47.0_{-6.3}^{+6.3}$ & $41.7^{+3.8}_{-4.1}$ & $41.0_{-3.4}^{+3.5}$ & $45.3_{-4.8}^{+4.1}$ & $42.6_{-4.9}^{+5.8}$ & $45.5^{+4.9}_{-4.9}$ \\
    \hline

    \begin{tabular}[c]{@{}c@{}}$q=0.68$\\ \citep{2017MNRAS.472.2937H} \end{tabular} & $44.6_{-4.0}^{+4.8}$ & $43.1_{-3.6}^{+4.4}$ & $48.8_{-7.1}^{+7.5}$ & $43.0^{+4.5}_{-4.4}$ & $42.2_{-3.6}^{+4.4}$ & $46.5_{-5.5}^{+4.9}$ & $44.0_{-5.3}^{+5.3}$ & $46.8^{+5.6}_{-5.2}$ \\

    \hline

   Bootstrapping & $47.1_{-1.5}^{+1.5}$ & $45.6_{-1.3}^{+1.3}$ & $44.8_{-2.0}^{+2.4}$ & - & $44.0_{-1.3}^{+1.4}$ & $43.0_{-1.3}^{+1.4}$ & $40.3_{-2.7}^{+2.5}$ & - \\

    \hline
\end{tabular}
\end{adjustbox}
\caption{Resumen de los valores de $i$ destacados, obtenidos mediante diferentes métodos.}
\label{resultados finales}
\end{table}

Answer 1

I endorse Mico's idea, but I'd make a few changes.

First off, a numeric table always benefits from siunitx; in this case there are two main reasons:

1. the decimal comma is used throughout without the need of trusting babel-spanish;

2. the asymmetric uncertainties are input in a much easier and uniform way.

The other main change is to set the citation callouts in a legend below the table, saving precious space. Finally, I got rid of the abbreviations and of unnecessary horizontal rules.

\documentclass[12pt,a4paper]{article}
 
%\usepackage[utf8]{inputenc}  % that's the default nowadays
\usepackage[top=3cm,bottom=1.8cm,hmargin=2cm]{geometry}

\usepackage[T1]{fontenc}
\usepackage[spanish,es-tabla]{babel}

\usepackage[authoryear,round]{natbib} % for \citep macro

\usepackage{booktabs}        % for well-spaced horizontal rules
\usepackage{siunitx}         % for numeric tables

\begin{document}

\begin{table}[htp]
\setlength\tabcolsep{0pt}

\begin{tabular*}{\textwidth}{
  @{\extracolsep{\fill}}
  l
  *{8}{S[table-format=2.1(1)]}
  @{}
}
\toprule
& \multicolumn{8}{c}{Filtros} \\
\cmidrule{2-9}
& \multicolumn{4}{c}{Conjunctos de datos original}  
& \multicolumn{4}{c}{Conjunctos de datos reducido} \\
\cmidrule{2-5} \cmidrule{6-9}
& {I} & {R} & {V} & {I,\,R,\,V} & {I} & {R} & {V} & {I,\,R,\,V} \\
\midrule

$q=0.62$ (a)
& 43.2(42:47) & 41.9(36:34) & 47.0(63:63) & 41.7(38:41) 
& 41.0(35:34) & 45.3(41:48) & 42.6(58:49) & 45.5(40:49) \\

\addlinespace

$q=0.68$ (b)
& 44.6(48:40) & 43.1(44:36) & 48.8(75:71) & 43.0(45:44)
& 42.2(44:36) & 46.5(49:55) & 44.0(53:53) & 46.8(56:52) \\

\addlinespace

Bootstrapping
& 47.1(15:15) & 45.6(13:13) & 44.8(24:20) & {--} 
& 44.0(14:13) & 43.0(14:13) & 40.3(25:27) & {--} \\

\midrule[\heavyrulewidth]

\multicolumn{9}{l}{%
  (a) \citet{2007ApJ...662..564B},\quad
  (b) \citet{2017MNRAS.472.2937H}
}
\end{tabular*}

\caption{Resumen de los valores de i destacados, obtenidos mediante diferentes métodos.}
\label{resultados_finales}

\end{table}

\end{document}

Production note: I used dummy bibliographic data to match the citations in your picture.

Answer 2

It suffices to (a) switch to a tabular* environment (with a target width of \textwidth) while setting \tabcolsep -- which governs the amount of intercolumn whitespace padding -- to 0pt and (b) shorten some of the header cells. I would also get rid of all vertical rules and replace the basic \hline directives with the line-drawing macros of the booktabs package. Your readers may appreciate the table's more open and inviting "look".

\documentclass[12pt,a4paper]{article}
 
%\usepackage[utf8]{inputenc}  % that's the default nowadays
\usepackage[top=3cm,bottom=1.8cm,hmargin=2cm]{geometry}

\usepackage[T1]{fontenc}
\usepackage[spanish,es-tabla]{babel}

\usepackage{array} % for \newcolumntype macro
\newcolumntype{C}{>{$}c<{$}} % automatic math mode

\usepackage[numbers]{natbib} % for \citep macro
\usepackage{booktabs}        % for well-spaced horizontal rules
\newcommand\mytab[1]{\begin{tabular}{l} #1 \end{tabular}} % handy shortcut macro

\begin{document}

\begin{table}
\setlength\tabcolsep{0pt}

\begin{tabular*}{\textwidth}{@{\extracolsep{\fill}} l *{8}{C} }
\toprule
& \multicolumn{4}{c}{Filtros, Conj.\ de datos original}  
& \multicolumn{4}{c}{Filtros, Conj.\ de datos reducido} \\
\cmidrule(l){2-5} \cmidrule(l){6-9}
& I & R & V & I,R,V & I & R & V & I,R,V \\
\midrule

\mytab{$q{=}0.62$\\ \citep{2007ApJ...662..564B}}
& 43.2_{-4.7}^{+4.2} & 41.9_{-3.4}^{+3.6} 
& 47.0_{-6.3}^{+6.3} & 41.7^{+3.8}_{-4.1} 
& 41.0_{-3.4}^{+3.5} & 45.3_{-4.8}^{+4.1} 
& 42.6_{-4.9}^{+5.8} & 45.5^{+4.9}_{-4.9} \\

\midrule

\mytab{$q{=}0.68$\\ \citep{2017MNRAS.472.2937H}}
& 44.6_{-4.0}^{+4.8} & 43.1_{-3.6}^{+4.4} 
& 48.8_{-7.1}^{+7.5} & 43.0^{+4.5}_{-4.4} 
& 42.2_{-3.6}^{+4.4} & 46.5_{-5.5}^{+4.9} 
& 44.0_{-5.3}^{+5.3} & 46.8^{+5.6}_{-5.2} \\

\midrule

Bootstr. 
& 47.1_{-1.5}^{+1.5} & 45.6_{-1.3}^{+1.3} 
& 44.8_{-2.0}^{+2.4} & \mbox{--} 
& 44.0_{-1.3}^{+1.4} & 43.0_{-1.3}^{+1.4} 
& 40.3_{-2.7}^{+2.5} & \mbox{--} \\

\bottomrule
\end{tabular*}

\caption{Resumen de los valores de i destacados, obtenidos mediante diferentes métodos.}
\label{resultados_finales}
\end{table}

\end{document}

Answer 3

Consider swapping your layout from horizontal (current) to vertical:

\documentclass{article}

\usepackage{makecell,booktabs}

\begin{document}

\begin{figure}
  \centering
  \begin{tabular}{ l l c c c }
    \toprule
    & \multicolumn{1}{c}{Filtro(s)} &
      \makecell[b]{$q = 0.62$ \\ (Bitner et al.~2007)} & % \citep{2007ApJ...662..564B}
      \makecell[b]{$q = 0.68$ \\ (Hill et al.~2017)} & %\citep{2017MNRAS.472.2937H}
      Bootstrapping \\
    \midrule
    \multicolumn{5}{l}{\itshape Conjunto de datos original} \\[0.5em]
    & I       & $43.2_{-4.7}^{+4.2}$ & $44.6_{-4.0}^{+4.8}$ & $47.1_{-1.5}^{+1.5}$ \\[0.5em]
    & R       & $41.9_{-3.4}^{+3.6}$ & $43.1_{-3.6}^{+4.4}$ & $45.6_{-1.3}^{+1.3}$ \\[0.5em]
    & V       & $47.0_{-6.3}^{+6.3}$ & $48.8_{-7.1}^{+7.5}$ & $44.8_{-2.0}^{+2.4}$ \\[0.5em]
    & I, R, V & $41.7^{+3.8}_{-4.1}$ & $43.0^{+4.5}_{-4.4}$ &          --          \\
    \addlinespace
    \multicolumn{5}{l}{\itshape Conjunto de datos reducido} \\[0.5em]
    & I       & $44.0_{-1.3}^{+1.4}$ & $41.0_{-3.4}^{+3.5}$ & $42.2_{-3.6}^{+4.4}$ \\[0.5em]
    & R       & $43.0_{-1.3}^{+1.4}$ & $45.3_{-4.8}^{+4.1}$ & $46.5_{-5.5}^{+4.9}$ \\[0.5em]
    & V       & $42.6_{-4.9}^{+5.8}$ & $44.0_{-5.3}^{+5.3}$ & $40.3_{-2.7}^{+2.5}$ \\[0.5em]
    & I, R, V & $45.5^{+4.9}_{-4.9}$ & $46.8^{+5.6}_{-5.2}$ &          --          \\
    \bottomrule
  \end{tabular}
  \caption{A figure caption}
\end{figure}

\end{document}

Answer 4

I endorse combination of the Werner's and egreg's ideas, but with use of ˙tabularray` package:

\documentclass[12pt,a4paper]{article}
\usepackage{tabularray}
\UseTblrLibrary{booktabs,               % for well-spaced horizontal rules
                siunitx}                % for numeric tables

\begin{document}
    \begin{table}[htp]
    \centering
\begin{tblr}{colsep = 10pt, 
             colspec= {@{} ll  *{3}{Q[c, si={table-format=2.1(2)}]} @{}},
             row{1} = {guard, b},
             rowsep = 3pt
            }
    \toprule
    &   Filtro(s)
        &   {$q = 0.62$ \\ (Bitner et al.~2007)} 
            &   {$q = 0.68$ \\ (Hill et al.~2017)} 
                &   Bootstrapping                           \\
    \midrule
\SetCell[c=4]{font=\itshape}  Conjunto de datos original
    &   &   &   &   &                                       \\
    & I         & 43.2(4.2:4.7) & 44.6(4.0:4.8)   & 47.0(1.5:1.5) \\
    & R         & 41.9(3.4:3.6) & 43.1(3.6:4.4)   & 45.6(1.3:1.3) \\
    & V         & 47.0(6.3:6.3) & 48.8(7.1:7.5)   & 44.8(2.0:2.4) \\
    & I, R, V   & 41.7(3.8:4.1) & 43.0(4.5:4.4)   &   {{{--}}}    \\
    \addlinespace
\SetCell[c=4]{font=\itshape}  Conjunto de datos reducido
    &   &   &   &   &                                  \\
    & I         & 44.0(1.3:1.4) & 41.0(3.4:3.5) &  42.2(3.6:4.4)     \\
    & R         & 43.0(1.3:1.4) & 45.3(4.8:4.1) &  46.5(5.5:4.9)     \\
    & V         & 42.6(4.9:5.8) & 44.0(5.3:5.3) &  40.3(2.7:2.5)     \\
    & I, R, V   & 45.5(4.9:4.9) & 46.8(5.6:5.2) &  {{{--}}}        \\
    \bottomrule
\end{tblr}

\caption{Resumen de los valores de i destacados, obtenidos mediante diferentes métodos.}
\label{resultados_finales}
    \end{table}
\end{document}