**My Experiments with R » R**, and kindly contributed to R-bloggers)

I use Sweave regularly for most of my writing and love the way it works. However, one issue that often irks me is the inability to globally set the number of digits to display. Here is a minimal example that illustrates my point.

<>= options(digits = 2); x = 1.2345; y = 1.214432532; z = 124.23414513; @

If we now display the numbers using Sexpr, this is what we get

x = 1.2345

y = 1.214432532

z = 124.23414513

Note how setting the number of digits to display using `options`

did not have any impact. The reason for this is that `Sexpr`

treats every variable as a character and hence the digits option does not work.

A simple solution to rectify this issue is to use functions like `format`

or `round`

to preset the number of digits inside the sweave chunk. However, this approach requires one to apply these functions on every variable individually, which in my opinion leads to some ugly code within the chunks.

This set me thinking whether it was possuble to write a function that (a) selects all the numeric variables created inside a chunk, and (b) formats them using global options. It turns out that it is possible to do this and in fact is quite straightforward.

<>= format_all = function(...){ library(plyr) # get all numeric variables num.obj = ls.str(mode = 'numeric', envir = .GlobalEnv); # apply format to all numeric variables l_ply(num.obj, function(.x) assign(.x, as.numeric(format(get(.x), ...)), envir = .GlobalEnv)) rm(tmp); } format_all(digits = 3, nsmall = 2); @

If we now display the numbers using `Sexpr`

, this is what we get

x = 1.23

y = 1.21

z = 124.23

**leave a comment**for the author, please follow the link and comment on their blog:

**My Experiments with R » R**.

R-bloggers.com offers

**daily e-mail updates**about R news and tutorials on topics such as: Data science, Big Data, R jobs, visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...