I'm trying to create a standard regression table with standard errors in parentheses below the coefficients, while using siunitx to align the numbers on the decimal point, and to round the numbers.
I use the option input-symbols=() to make siunitx ignore the parentheses, which works fine for alignment, but the numbers in parentheses are not rounded:
\documentclass{article}
\usepackage{siunitx}
\begin{document}
\begin{tabular}{l*{2}{S[input-symbols=(),round-mode=figures,round-precision=1]}}
\hline
& \multicolumn{1}{c}{(1)} & \multicolumn{1}{c}{(2)} \\
\hline
X & 0.0131 & 0.00265{***} \\
& (0.000731) & (0.000547) \\
\hline
\end{tabular}
\end{document}
If I instead put the parentheses in curly brackets, the rounding works, but the left paranthesis ends up left aligned in the table cell:
\documentclass{article}
\usepackage{siunitx}
\begin{document}
\begin{tabular}{l*{2}{S[round-mode=figures,round-precision=1]}}
\hline
& \multicolumn{1}{c}{(1)} & \multicolumn{1}{c}{(2)} \\
\hline
X & 0.0131 & 0.00265{***} \\
& {(}0.000731{)} & {(}0.000547{)} \\
\hline
\end{tabular}
\end{document}
Is there a way to get rounding to work with parantheses? I have the latest version of siunitx, 2.4j.
table-align-text-pre=false,table-align-text-post=false,table-space-text-pre=(,table-space-text-post=)and wrapping all parentheses in curly brackets gives me the output I want.Still, it would be very nice if this was possible without the curly brackets and the extra options.
– standard_error Mar 20 '12 at 12:50