Using btabler

library(btabler)
#> Loading required package: xtable

btabler is a wrapper for the xtable package which adds some new functionality for merging headers, adding footers etc.

To demonstrate how btabler

df <- data.frame(name = c("", "Row 1", "Row2"),
                 out_t = c("Total", "t1", "t1"),
                 out_1 = c("Group 1", "g11", "g12"), 
                 out_2 = c("Group 2", "g21", "g22"))

btable(df, nhead = 1, nfoot = 0, caption = "Table1")
#> \begingroup\fontsize{8pt}{12pt}\selectfont
#> \begin{longtable}{lccc}
#> \caption{Table1} \\ 
#>   \toprule
#>  \multicolumn{1}{l}{} & \multicolumn{1}{c}{Total} & \multicolumn{1}{c}{Group 1} & \multicolumn{1}{c}{Group 2} \\ 
#>    \hline 
#>  \endhead 
#>  \hline 
#> \multicolumn{4}{l}{\textit{continued on next page}} 
#>  \endfoot 
#>  \endlastfoot 
#> Row 1 & t1 & g11 & g21 \\ 
#>   Row2 & t1 & g12 & g22 \\ 
#>    \bottomrule
#> \end{longtable}
#> \endgroup

In the compiled PDF, this looks substantially better of course…

Templates

btabler requires that a few specific LaTeX packages are included in the header of your Rmd or Rnw/`tex’ file:

# .tex
\usepackage{longtable}
\usepackage{booktabs}
\usepackage{float}
\usepackage{array}

# .Rmd
header-includes:
  - \usepackage{longtable}
  - \usepackage{booktabs}
  - \usepackage{float}
  - \usepackage{array}

For convenience, btabler provides templates for each which can be accessed via use_btabletemplate:

use_btabletemplate("report")             # creates report.Rmd in the working directory
use_btabletemplate("report", "Rnw")      # creates report.Rnw in the working directory
use_btabletemplate("code/report", "Rnw") # creates report.Rnw in the code directory

Headers and footers

Where there are multiple header lines, the nhead argument can be used and any neighboring cells in those first rows will be merged.

df <- data.frame(name = c("", "", "Row 1", "Row2"),
                 out_t = c("Total", "mean (sd)", "t1", "t1"),
                 out_1 = c("Group 1", "mean (sd)", "g11", "g12"),
                 out_2 = c("Group 2", "mean (sd)", "g21", "g22"))
btable(df, nhead = 2, nfoot = 0, caption = "Table1")

btable italicizes the second row of the header by default. This can be changed via the head_it argument:

btable(df, nhead = 2, nfoot = 0, caption = "Table1", 
       head_it = NA)

Likewise, aggregation of the header can also be turned of

btable(df, nhead = 2, nfoot = 0, caption = "Table1", 
       head_it = NA, aggregate = FALSE)

Footers included in the dataframe can also be added:

df <- data.frame(name = c("", "Row 1", "Row2", "*Footer"),
                 out_t = c("Total", "t1", "t1", ""),
                 out_1 = c("Group 1", "g11", "g12", ""),
                 out_2 = c("Group 2", "g21", "g22", ""))
btable(df, nhead=1, nfoot=1, caption="Table1", 
       rulelength = "6cm")

Alignment

Alignment can be changed via the aligntot argument. For example, we could specify that the first column be left aligned and all other columns should be centered aligned:

btable(df, nhead = 1, nfoot = 0, caption = "Table1", aligntot = "lccc")

Custom column types

It’s possible to create new column types in LaTeX and use them in btabler.

The following creates a new column type if put in the LaTeX or Rmd header

# .tex
\newcolumntype{P}[1]{>{\centering\arraybackslash}p{#1}}

# .Rmd
header-includes:
  ... # other requirements
  - \newcolumntype{P}[1]{>{\centering\arraybackslash}p{#1}}

This can then be used in btable in the aligntot argument (note that xtable warns about non-standard, adding warning = FALSE to the chunk header might be useful…)

btable(df, nhead = 1, nfoot = 0, 
       caption = "Table1", 
       aligntot = "P{3cm}P{1.5cm}P{1.5cm}P{1.5cm}")