I'm about to write a function that yields both a screen output and a LaTeX/PDF-output of lists of regression tables and have problems with the decimal point alignment of the latter.
From this toy table regmat,
1 2 3 4
(Intercept) " 33.99 ***" " 3370.83 ***" " 39.42 ***" " 46.80 ***"
" (1.89) " " (260.49) " " (2.64) " " (6.37) "
cyl " --- " " --- " " -1.51 ** " " --- "
" --- " " --- " " (0.42) " " --- "
factor(am)1 " --- " " 180.92 " " 0.18 " " 0.19 "
" --- " " (139.63) " " (1.30) " " (1.49) "
factor(cyl)6 " -4.26 ** " " -303.13 * " " --- " " -1.28 "
" (1.39) " " (140.73) " " --- " " (1.53) "
factor(cyl)8 " -6.07 ***" " -216.37 " " --- " " -1.39 "
" (1.65) " " (228.43) " " --- " " (2.16) "
hp " --- " " -3.21 * " " --- " " -0.10 ** "
" --- " " (1.37) " " --- " " (0.03) "
wt " -3.21 ***" " -249.68 ** " " -3.13 ** " " -7.12 ** "
" (0.75) " " (88.56) " " (0.91) " " (2.24) "
wt:hp " --- " " --- " " --- " " 0.02 * "
" --- " " --- " " --- " " (0.01) "
Deviance " 1126.05 " "11260471.88 " " 1126.05 " " 1126.05 "
Df " 31 " " 31 " " 31 " " 31 "
Log-Likelihood " -73.31 " " -217.60 " " -73.99 " " -67.34 "
AIC " 156.62 " " 449.20 " " 157.99 " " 150.69 "
BIC " 163.95 " " 459.46 " " 165.32 " " 162.41 "
Resid. deviance " 183.06 " " 1510255.92 " " 191.05 " " 126.07 "
Resid. Df " 28 " " 26 " " 28 " " 25 "
Num. obs. " 32 " " 32 " " 32 " " 32 "
I get a nice screen output with as.data.frame(regmat);
> as.data.frame(regmat)
1 2 3 4
(Intercept) 33.99 *** 3370.83 *** 39.42 *** 46.80 ***
(1.89) (260.49) (2.64) (6.37)
cyl --- --- -1.51 ** ---
--- --- (0.42) ---
factor(am)1 --- 180.92 0.18 0.19
--- (139.63) (1.30) (1.49)
factor(cyl)6 -4.26 ** -303.13 * --- -1.28
(1.39) (140.73) --- (1.53)
factor(cyl)8 -6.07 *** -216.37 --- -1.39
(1.65) (228.43) --- (2.16)
hp --- -3.21 * --- -0.10 **
--- (1.37) --- (0.03)
wt -3.21 *** -249.68 ** -3.13 ** -7.12 **
(0.75) (88.56) (0.91) (2.24)
wt:hp --- --- --- 0.02 *
--- --- --- (0.01)
Deviance 1126.05 11260471.88 1126.05 1126.05
Df 31 31 31 31
Log-Likelihood -73.31 -217.60 -73.99 -67.34
AIC 156.62 449.20 157.99 150.69
BIC 163.95 459.46 165.32 162.41
Resid. deviance 183.06 1510255.92 191.05 126.07
Resid. Df 28 26 28 25
Num. obs. 32 32 32 32
the decimal points are all aligned, even in lines with parentheses, and the integers in the lower GOF section end before the points.
Now I want a similar LaTeX/PDF output with xtable using rmarkdown and this code,
library(xtable)
xtbl <- xtable(regmat, comment=FALSE)
align(xtbl) <- c("L{2.4cm}", rep(c("L{2cm}@{\\hskip 0.1cm}"), 4))
print(xtbl,
tabular.environment="tabularx",
width=".5\\textwidth",
hline.after=c(-1, 0, 16, nrow(regmat)),
comment=FALSE)
where the YAML looks like this
---
output:
pdf_document:
includes:
in_header: "X:/<myDir>/header.tex"
classoption:
- a4paper
- landscape
geometry: margin=2cm
header-includes:
- \pagenumbering{gobble}
- \usepackage{dcolumn}
---
and header.tex -- as suggested here -- looks like so,
\usepackage{array}
\usepackage{tabularx}
\newcolumntype{L}[1]{>{\raggedright\let\newline\\
\arraybackslash\hspace{0pt}}m{#1}}
\newcolumntype{C}[1]{>{\centering\let\newline\\
\arraybackslash\hspace{0pt}}m{#1}}
\newcolumntype{R}[1]{>{\raggedleft\let\newline\\
\arraybackslash\hspace{0pt}}m{#1}}
\newcolumntype{P}[1]{>{\raggedright\tabularxbackslash}p{#1}}
obtaining this failed LaTeX/PDF-table result:
Looks not bad, but as one can see the left-alignment starts at the first non-space character and all the white spaces seem to be ignored.
In rep(c("L{2cm}@{\\hskip 0.1cm}") I tried all of L, C, R, and P of the header.tex, but none of them aligns the values at the decimal point, i.e. doesn't seem to use the package \usepackage{dcolumn} which is written to align decimal points.
I don't care about dcolumn, though, I just find it weird that my white spaces are ignored.
How can I get the same alignment in my LaTeX-PDF-table as in the screen output shown above, either with dcolumn or by not ignoring the white spaces?
Data
regmat <- structure(c(" 33.99 ***", " (1.89) ", " --- ",
" --- ", " --- ", " --- ", " -4.26 ** ",
" (1.39) ", " -6.07 ***", " (1.65) ", " --- ",
" --- ", " -3.21 ***", " (0.75) ", " --- ",
" --- ", " 1126.05 ", " 31 ", " -73.31 ",
" 156.62 ", " 163.95 ", " 183.06 ", " 28 ",
" 32 ", " 3370.83 ***", " (260.49) ", " --- ",
" --- ", " 180.92 ", " (139.63) ", " -303.13 * ",
" (140.73) ", " -216.37 ", " (228.43) ", " -3.21 * ",
" (1.37) ", " -249.68 ** ", " (88.56) ", " --- ",
" --- ", "11260471.88 ", " 31 ", " -217.60 ",
" 449.20 ", " 459.46 ", " 1510255.92 ", " 26 ",
" 32 ", " 39.42 ***", " (2.64) ", " -1.51 ** ",
" (0.42) ", " 0.18 ", " (1.30) ", " --- ",
" --- ", " --- ", " --- ", " --- ",
" --- ", " -3.13 ** ", " (0.91) ", " --- ",
" --- ", " 1126.05 ", " 31 ", " -73.99 ",
" 157.99 ", " 165.32 ", " 191.05 ", " 28 ",
" 32 ", " 46.80 ***", " (6.37) ", " --- ",
" --- ", " 0.19 ", " (1.49) ", " -1.28 ",
" (1.53) ", " -1.39 ", " (2.16) ", " -0.10 ** ",
" (0.03) ", " -7.12 ** ", " (2.24) ", " 0.02 * ",
" (0.01) ", " 1126.05 ", " 31 ", " -67.34 ",
" 150.69 ", " 162.41 ", " 126.07 ", " 25 ",
" 32 "), .Dim = c(24L, 4L), .Dimnames = list(c("(Intercept)",
" ", "cyl", " ", "factor(am)1", " ", "factor(cyl)6", " ",
"factor(cyl)8", " ", "hp", " ", "wt", " ",
"wt:hp", " ", "Deviance", "Df", "Log-Likelihood",
"AIC", "BIC", "Resid. deviance", "Resid. Df", "Num. obs."), c("1",
"2", "3", "4")))
