我正在尝试使用 tabularx 在回归表上实现小数对齐。我使用 dcolumn 包来实现这一点，但存在两个问题：(i) 三列中只有两列是小数对齐的，(ii) 有一列没有对齐整数变量，在我的情况下是观察次数。有人能帮我吗？我尝试过将 siunitx 包与 tabularx 一起使用，但没有成功。
欢迎所有解决方案！
谢谢。
代码：
\documentclass[12pt]{article}
\usepackage{bigstrut}
\usepackage{graphics}
\usepackage{endnotes}
\usepackage{setspace}
\usepackage{epsfig}
\usepackage{amssymb}
\usepackage{amsfonts}
\usepackage{makeidx}
\usepackage{verbatim}
\usepackage{bm}
\usepackage{latexsym}
\usepackage[latin1]{inputenc}
\usepackage[T1]{fontenc} % Font output type
\usepackage{ae,aecompl}
\usepackage{amsthm}
\usepackage{amsmath}
\usepackage{bbm}
\usepackage{rotating}
\usepackage{graphicx}
\usepackage{dsfont}
\usepackage{epstopdf}
\usepackage{url}
\usepackage{array}
\usepackage{placeins}
\usepackage{floatrow}
\usepackage{datetime}
\usepackage{tabularx} % provides a column type called "X" that should satisfy your professed need to have several equal-width columns
\usepackage{booktabs}
\usepackage{threeparttable}
\usdate
%\usepackage{rotfloat}
\usepackage{rotating}
\usepackage{enumerate}
\usepackage{colortbl}
%\usepackage[pdftex]{color,graphicx}
\usepackage{lscape}
\usepackage{longtable}
\usepackage{float}
\floatstyle{plaintop}
\restylefloat{table}
\usepackage[labelfont=bf, labelsep=period, justification=centering, font={small,doublespacing}, skip=0pt]{caption}%define table values
%\usepackage[labelfont=bf, labelsep=period, center, scriptsize]{caption}
\usepackage[table]{xcolor}
%\usepackage{indentfirst}
%\bibliography{refs}
\usepackage{hyper}
%\usepackage{pgf}
\usepackage{multirow}
\usepackage{pdflscape}
%\usepackage{bibtex}
\usepackage[comma, sort&compress, longnamesfirst]{natbib}
\usepackage[flushmargin,hang]{footmisc}
\setlength{\footnotemargin}{0.5em} % just to show clearly equal output
\usepackage{titlesec}
\usepackage{siunitx}
\usepackage{dcolumn}
\newcolumntype{L}{>{\raggedleft\arraybackslash}X} % handy shortcut macro
\newcolumntype{R}{>{\raggedright\arraybackslash}X} % handy shortcut macro
\newcolumntype{C}{>{\centering\arraybackslash}X} % handy shortcut macro
\newcolumntype{d}[1]{D{.}{\cdot}{#1}} % and yet another handy shortcut macro
\begin{document}
\begin{table}[h]
\renewcommand{\arraystretch}{1.2} %Vertically expands the table
%\addtolength{\tabcolsep}{5pt}%Horizontally expands the table
\begin{threeparttable}
%\begin{small}
\centering
\def\sym#1{\ifmmode^{#1}\else\(^{#1}\)\fi}
\caption{In-sample forecasting monthly excess stock returns ($R_{i,t+1}-R_{f,t+1}$).}
%\begin{tabular}{lccccccc}
\begin{tabularx}{\textwidth}{l l*{3}{d{-2}}}% the tabularx package gives you: the total width as a first parameter, and a new column type X, all X columns will grow to fill up the total width.
\toprule
Dependent Variable: $R_{i,t+1} - R_{f,t+1}$ & \multicolumn{1}{r}{(1)} & (2) & (3) \\
\midrule
Constant & -0.001 & -0.001 & 0.001 \\
& (-0.10) & (-0.09) & (0.16) \\
$R_{m,t} - R_{f,t}$ & 0.111\sym{*} & 0.122\sym{**}& 0.113\sym{*} \\
& (1.92) & (2.09) & (1.91) \\
$X$ & $-1.301^{*}$ & -1.296\sym{*} & -1.308\sym{*} \\
& (-1.92) & (-1.91) & (-1.95) \\
$Y$ & 5.900\sym{**}& 5.889\sym{**}& 5.357\sym{**}\\
& (2.24) & (2.23) & (2.01) \\
$Z$ & 0.304 & 0.232 & 2.318 \\
& (0.03) & (0.03) & (0.26) \\
$W$ & & -0.001\sym{**}& -0.001\sym{**}\\
& & (-2.44) & (-2.70) \\
$A$ & & & -0.056 \\
& & & (-0.45) \\
N obs & 8981 & \multicolumn{1}{c}{8981} & 8991 \\
$R^2$ & 0.014 & 0.015 & 0.015 \\
\bottomrule
%\end{tabular}%
\end{tabularx}%
\begin{tablenotes}[flushleft]%[para]
\scriptsize
\item
\item \sym{*} Scaled by 100.
\end{tablenotes}
\label{panel reg_full sample}%
%\end{small}
\end{threeparttable}
\end{table}
\end{document}
答案1
请尝试将示例修剪为仅需要的包，在这种情况下，您的示例只需要
\documentclass[12pt]{article}
\usepackage{tabularx} % provides a column type called "X" that should satisfy your professed need to have several equal-width columns
\usepackage{booktabs}
\usepackage{threeparttable}
%\usdate
\usepackage{dcolumn}
\newcolumntype{d}[1]{D{.}{\cdot}{#1}} % and yet another handy shortcut macro
\begin{document}
您已指定 5 列，即前两列l和接下来的三列d，但您只有 4 列数据，因此第三列d从未使用过，这解释了为什么您只能在两列中获得小数对齐。
tabularx通过改变宽度来工作换行在X列中没有换行符，也没有 X，因此无法按预期工作。只需使用tabular（如果您真的想要过度分隔可以使用的列tabular*，但这只会使表格更难读取）
这里我将 8981 中的两个对齐为整数，但我认为将列 (2) 居中放置会更好。
\documentclass[12pt]{article}
\usepackage{booktabs}
\usepackage{threeparttable}
%\usdate
\usepackage{dcolumn}
\newcolumntype{d}[1]{D{.}{\cdot}{#1}} % and yet another handy shortcut macro
\begin{document}
\begin{table}[h]
\renewcommand{\arraystretch}{1.2} %Vertically expands the table
%\addtolength{\tabcolsep}{5pt}%Horizontally expands the table
\begin{threeparttable}
%\begin{small}
\centering
\def\sym#1{\ifmmode^{#1}\else\(^{#1}\)\fi}
\caption{In-sample forecasting monthly excess stock returns ($R_{i,t+1}-R_{f,t+1}$).}
\begin{tabular}{l *{3}{d{-2}}}% the tabularx package gives you: the total width as a first parameter, and a new column type X, all X columns will grow to fill up the total width.
\toprule
Dependent Variable: $R_{i,t+1} - R_{f,t+1}$ & \multicolumn{1}{c}{(1)} &
\multicolumn{1}{c}{(2)} &
\multicolumn{1}{c}{(3)} \\
\midrule
Constant & -0.001 & -0.001 & 0.001 \\
& (-0.10) & (-0.09) & (0.16) \\
$R_{m,t} - R_{f,t}$ & 0.111\sym{*} & 0.122\sym{**}& 0.113\sym{*} \\
& (1.92) & (2.09) & (1.91) \\
$X$ & -1.301^{*} & -1.296\sym{*} & -1.308\sym{*} \\
& (-1.92) & (-1.91) & (-1.95) \\
$Y$ & 5.900\sym{**}& 5.889\sym{**}& 5.357\sym{**}\\
& (2.24) & (2.23) & (2.01) \\
$Z$ & 0.304 & 0.232 & 2.318 \\
& (0.03) & (0.03) & (0.26) \\
$W$ & & -0.001\sym{**}& -0.001\sym{**}\\
& & (-2.44) & (-2.70) \\
$A$ & & & -0.056 \\
& & & (-0.45) \\
$N$ obs & 8981 & \multicolumn{1}{c}{8981} & 8991 \\
$R^2$ & 0.014 & 0.015 & 0.015 \\
\bottomrule
\end{tabular}%
\begin{tablenotes}[flushleft]%[para]
\scriptsize
\item
\item \sym{*} Scaled by 100.
\end{tablenotes}
\label{panel reg_full sample}%
%\end{small}
\end{threeparttable}
\end{table}
\end{document}