I produced a cross table with the CrossTable command from the gmodels package such as:
library(gmodels)
library(descr)
a<-CrossTable(mtcars$cyl,mtcars$gear,prop.chisq = FALSE)
If I create a table for LaTex using xtable, the frequencies for row totals (last column) and column totals (last row) are missing. However, the total counts are there.
I used:
print(xtable(a,digits=3))
to achieve this. How do I get these values in as well?
Thanks for your support.
EDIT:
Here an example of the output:
a
===========================================
mtcars$gear
mtcars$cyl 3 4 5 Total
-------------------------------------------
4 1 8 2 11
0.091 0.727 0.182 **0.344**
0.067 0.667 0.400
0.031 0.250 0.062
-------------------------------------------
6 2 4 1 7
0.286 0.571 0.143 **0.219**
0.133 0.333 0.200
0.062 0.125 0.031
-------------------------------------------
8 12 0 2 14
0.857 0.000 0.143 **0.438**
0.800 0.000 0.400
0.375 0.000 0.062
-------------------------------------------
Total 15 12 5 32
**0.469** **0.375** **0.156**
===========================================
I highlighted the missing values after xtable conversion with *. But when I have it converted to Tex with print it looks like:
\begin{table}[ht]
\centering
\begin{tabular}{llrrrr}
\hline
& mtcars\$cyl & 3 & 4 & 5 & Total \\
\hline
1 & 4 & 1 & 8 & 2 & 11 \\
2 & & 0.091 & 0.727 & 0.182 & \\
3 & & 0.067 & 0.667 & 0.400 & \\
4 & & 0.031 & 0.250 & 0.062 & \\
5 & 6 & 2 & 4 & 1 & 7 \\
6 & & 0.286 & 0.571 & 0.143 & \\
7 & & 0.133 & 0.333 & 0.200 & \\
8 & & 0.062 & 0.125 & 0.031 & \\
9 & 8 & 12 & 0 & 2 & 14 \\
10 & & 0.857 & 0.000 & 0.143 & \\
11 & & 0.800 & 0.000 & 0.400 & \\
12 & & 0.375 & 0.000 & 0.062 & \\
13 & Total & 15 & 12 & 5 & 32 \\
\hline
\end{tabular}
\end{table}
This could be viewed as a bug in the xtable.Crosstable method, which is in the descr package
environment(getS3method("xtable", "CrossTable"))
<environment: namespace:descr>
So you might like to contact the package maintainer about this. Meanwhile here is a hack that avoids re-computing the proportions yourself.
First capture the printed output from xtable to a temporary file and read each line back in as a separate string:
file <- tempfile()
capture.output(print(xtable(a, digits=3), include.rownames = FALSE),
file = file)
xout <- readLines(file)
I have assumed you don't want the rownames in your table. Do the same for the printed version of the CrossTable:
capture.output(CrossTable(mtcars$cyl,mtcars$gear,prop.chisq = FALSE),
file = file)
out <- readLines(file)
Next extract the body of the printed CrossTable output and "LaTeXize" it
body <- out[11:29]
## replace whitespace by column delimiter
body <- gsub(" *", " & ", body)
## add end of line markup for table rows
body <- gsub("([^-])$", "\\1 \\\\\\\\", body)
## replace dashed lines with hline
body <- gsub("-+", "\\\\hline", body)
## replace $ with \$
body <- gsub("$", "\\$", body, fixed = TRUE)
Finally print the header and footer from the printed xtable output with the new body:
cat(xout[1:6], body, xout[22:24], sep = "\n")
Producing the output
% latex table generated in R 3.1.1 by xtable 1.7-4 package
% Tue Oct 14 12:17:07 2014
\begin{table}[ht]
\centering
\begin{tabular}{lrrrr}
\hline
mtcars\$cyl & 3 & 4 & 5 & Total \\
\hline
4 & 1 & 8 & 2 & 11 \\
& 0.091 & 0.727 & 0.182 & 0.344 \\
& 0.067 & 0.667 & 0.400 & \\
& 0.031 & 0.250 & 0.062 & \\
\hline
6 & 2 & 4 & 1 & 7 \\
& 0.286 & 0.571 & 0.143 & 0.219 \\
& 0.133 & 0.333 & 0.200 & \\
& 0.062 & 0.125 & 0.031 & \\
\hline
8 & 12 & 0 & 2 & 14 \\
& 0.857 & 0.000 & 0.143 & 0.438 \\
& 0.800 & 0.000 & 0.400 & \\
& 0.375 & 0.000 & 0.062 & \\
\hline
Total & 15 & 12 & 5 & 32 \\
& 0.469 & 0.375 & 0.156 \\
\hline
\end{tabular}
\end{table}